The DOI matches its Crossref record; 101 of the 104 references it cites were found in a registry or an index; none of the 27 sources its claims cite is checked for support yet.
The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation
Open accessAbstract
Background: To evaluate binary classifications and their confusion matrices, scientific researchers can employ several statistical rates, accordingly to the goal of the experiment they are investigating. Despite being a crucial issue in machine learning, no widespread consensus has been reached on a unified elective chosen measure yet. Accuracy and F 1 score computed on confusion matrices have been (and still are) among the most popular adopted metrics in binary classification tasks. However, these statistical measures can dangerously show overoptimistic inflated results, especially on imbalanced datasets. Results: The Matthews correlation coefficient (MCC), instead, is a more reliable statistical rate which produces a high score only if the prediction obtained good results in all of the four confusion matrix categories (true positives, false negatives, true negatives, and false positives), proportionally both to the size of positive elements and the size of negative elements in the dataset. Conclusions: In this article, we show how MCC produces a more informative and truthful score in evaluating binary classifications than accuracy and F 1 score, by first explaining the mathematical properties, and then the asset of MCC in six synthetic use cases and in a real genomics scenario. We believe that the Matthews correlation coefficient should be preferred to accuracy and F 1 score in evaluating binary classification tasks by all scientific communities.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
The paper argues that the Matthews correlation coefficient (MCC) is a more reliable metric than accuracy and F1 score for evaluating binary classifications, especially on imbalanced datasets, because MCC requires good performance on both positive and negative classes. It supports this through mathematical analysis, six synthetic use cases, and a real colon cancer gene expression experiment.
Key findings
- In six synthetic use cases, MCC was the only metric that correctly indicated the prediction problem in all six. For Use case A1 (91 positives, 9 negatives; TP=90, FN=1, TN=0, FP=9), accuracy=0.90 and F1=0.95, but MCC=−0.03. For Use case C2 (11 positives, 89 negatives; TP=2, FN=9, TN=88, FP=1), accuracy=0.90, F1=0.29, and MCC=+0.31.
- Correlation analysis over all possible confusion matrices showed MCC and accuracy were strongly correlated, while F1 correlated less than 0.8 with the other two measures. For N=1000, PCC(MCC,F1)=0.775982, PCC(MCC,accuracy)=0.921030, and PCC(accuracy,F1)=0.778652.
- In the colon cancer gene expression dataset (2,000 gene probesets, 62 patients: 22 healthy controls and 40 colon cancer; 35.48% negatives, 64.52% positives), rankings differed by metric. MCC ranked gradient boosting top (MCC=+0.55, TP rate=0.85, TN rate=0.69), while F1 and accuracy ranked k-NN top (F1=0.87, accuracy=0.80, TP rate=0.92, TN rate=0.52). Radial SVM had F1=0.75 and accuracy=0.67 but TN rate=0.40 and MCC=+0.29.
- MCC produces a high score only if the classifier correctly predicts the majority of positive and negative instances; F1 is independent of true negatives and is not symmetric for class swapping, while accuracy can be overoptimistic on imbalanced datasets.
Methods
- The paper defines accuracy, MCC, and F1 score mathematically, explores their relationships using Pearson correlation coefficient over all possible confusion matrices for N≥10 up to 1000, and plots all 21,084,251 possible confusion matrices for a 500-sample dataset on the MCC/F1 plane.
- Six synthetic use cases vary class balance and imbalance with specified confusion matrices to compare metric behavior.
- The real genomics experiment uses colon cancer gene expression data from Alon et al. via the plsgenomics R package, applying gradient boosting, decision tree, k-nearest neighbors, linear SVM, and radial SVM. Training/test splits were 80/20 or 60/20/20 with validation for hyperparameter grid search; experiments were repeated ten times and averaged for MCC, F1, accuracy, TP rate, and TN rate.
Limitations
- The paper states that additional comparisons with other rates such as Cohen’s Kappa, Cramér’s V, and K measure would have provided further information, but the authors focused on accuracy and F1 because they are more commonly used in biomedical machine learning studies.
- The paper notes that MCC can be undefined when a whole row or column of the confusion matrix is zero, and handles these cases via an approximation using an arbitrarily small value ε; it also cites situations where MCC can display large fluctuations due to imbalanced outcomes.
- Evident from the text: the real-world validation uses a single colon cancer gene expression dataset, the six use cases are synthetic, and Table 5 excludes standard deviations to keep the table simple.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures
![Fig. 2 Use case A1 — Positively imbalanced dataset. a Barplot representing accuracy, F1, and normalized Matthews correlation coefficient (normMCC = (MCC + 1) / 2), all in the [0, 1] interval, where 0 is the worst possible score and 1 is the best possible score, applied to the Use case A1 positively imbalanced dataset. b Pie chart representing the amounts of true positives (TP), false negatives (FN), true negatives (TN), and false positives (FP). c Pie chart representing the dataset balance, as the amounts of positive data instances and negative data instances](/paper-figures/988371a13eb649a5bff767f781052a4f/the-advantages-of-the-matthews-correlation-coefficient-mcc-over-f1-score-and-figure-2.jpg)
Verification
Claims the paper makes
29 Own interpretation · 28 Own result · 15 Not checked yet · 13 No basis given
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“A potential problem with MCC lies in the fact that MCC is undefined when a whole row or column of M is zero, as it happens in the previously cited case of the trivial majority classifier.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Notation and mathematical foundations
plainn = 4- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Second, F1 score is independent from the number of samples correctly classified as negative.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“she/he would notice good results (F1 score = 0.75 and accuracy = 0.67)”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For the Use case A1 (negatively imbalanced dataset), the machine learning classifier was unable to correctly predict negative data instances, and it therefore produced confusion matrices featuring few true negatives (TN).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“First, the top classifier changes when we consider the ranking based on MCC, F1 score, or accuracy.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“MCC = +0.31 in the [–1, +1] interval”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In the Use case B1 (balanced dataset), the machine learning method was unable to correctly predict negative data instances, and therefore produced a confusion matrix featuring few true negatives (TN).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Finally, in the last Use case C2 (positively imbalanced dataset), the prediction technique failed in predicting negative elements, and therefore its confusion matrix showed a low percentage of true negatives.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“F1 score = 0.29 in the [0, 1] interval”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the radial SVM has performed poorly on the true negatives (TN rate = 0.40), by correctly predicting less than half patients.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Probably not a substantial claim. 15
- The quote only partly states this claim. 5
- It may state two separate claims as one. 36
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Here, accuracy showed a high value: 0.90 in the [0, 1] interval.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The quote only partly states this claim.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In this case, all three statistical rates showed a low score which emphasized the deficiency in the prediction process (accuracy = 0.24, F1 score = 0.12, and MCC = −0.24).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It ranges in the interval [−1,+1], with extreme values –1 and +1 reached in case of perfect misclassification and perfect classification, respectively, while MCC=0 is the expected value for the coin tossing classifier.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Notation and mathematical foundations
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“F1 ranges in [0,1], where the minimum is reached for TP=0, that is, when all the positive samples are misclassified, and the maximum for FN=FP=0, that is for perfect classification.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Notation and mathematical foundations
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Like for the Use case A2, accuracy, F1 and MCC show average or low result scores (accuracy = 0.56, F1 score = 0.31, and MCC = +0.17), correctly informing you about the non-optimal performance of the prediction method (Additional file 3a).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“If M has only one non-zero entry, this means that all samples in the dataset belong to one class, and they are either all correctly (for TP≠0 or TN≠0) or incorrectly (for FP≠0 or FN≠0) classified. In this situations, MCC=1 for the former case and MCC=−1 for the latter case.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Notation and mathematical foundations
plaindirectSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“If the dataset is highly unbalanced, ni≈1, and thus the accuracy measure gives an unreliable estimation of the goodness of the classifier.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Notation and mathematical foundations
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Checking only F1, one would read a good value (0.66 in the [0, 1] interval), and would be overall satisfied about the prediction (Additional file 2a).”
- Also stated in: § Use cases · hedged
“The Matthews correlation coefficient, instead, by showing a score close to random guessing (+0.07 in the [–1, +1] interval) would be able to inform that the machine learning method has been on the wrong track. Also, it is worth noticing that accuracy would provide with an informative result in this case (0.52 in the [0, 1] interval).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one. The quote only partly states this claim.
- Also stated in: § Use cases · hedged
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, accuracy and of F1 showed high values in this case: accuracy = 0.90 and F1 score = 0.95, both close to the best possible value 1.00 in the [0, 1] interval (Fig. 2a).”
- Also stated in: § Use cases · hedged
“Instead, if one decided to take advantage of the Matthews correlation coefficient in the Use case A1, he/she would notice the resulting MCC = –0.03 (Fig. 2a).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one. The quote only partly states this claim.
- Also stated in: § Use cases · hedged
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Similar to the Use case A2 and B2, all three statistical scores generated low results that reflect the mediocre quality of the prediction: F1 score = 0.17 and accuracy = 0.10 in the [0, 1] interval, and MCC = −0.19 in the [–1, +1] interval (Additional file 4a).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“On the other hand, for a given value MCC=μ, the F1 score can range in [0,μ+1] if μ≤0 and in [μ,1] if μ>0, so that the width of the range is 1−|μ|, that is, it depends on the MCC value μ.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The two measures are reasonably concordant, but the scatterplot cloud is wide, implying that for each value of F1 score there is a corresponding range of values of MCC and vice versa, although with different width.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Overall, accuracy, F1, and MCC show reliable concordant scores for predictions that correctly classify both positives and negatives (having therefore many TP and TN), and for predictions that incorrectly classify both positives and negatives (having therefore few TP and TN); however, these measures show discordant behaviors when the prediction performs well just with one of the two binary classes.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“First, F1 varies for class swapping, while MCC is invariant if the positive class is renamed negative and vice versa.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Interestingly, the correlation grows with N, but the increments are limited.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“With these positions MCC is now defined for all confusion matrices M. As a consequences, MCC=0 for the trivial majority classifier, and 0 is also the expected value for the coin tossing classifier.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Notation and mathematical foundations
plainn = 4- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“MCC and accuracy resulted strongly correlated, while the Pearson coefficient is less than 0.8 for the correlation of F1 with the other two measures (Table 3).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
boostedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In fact, for any value F1=ϕ, the MCC varies approximately between [ϕ−1,ϕ], so that the width of the variability range is 1, independent from the value of ϕ.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
hedgedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“The micro/macro averaged F1 is invariant for class swapping and its behaviour is more similar to MCC.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Here, accuracy showed a high value: 0.90 in the [0, 1] interval.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“linear SVM is ranked forth in the MCC ranking and in the accuracy ranking, but ranked second in the F1 score ranking.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“k-NN, that is the top performing method both in the F1 score ranking and in the accuracy ranking, instead, obtained an excellent score for recall (TP rate = 0.92) but just sufficient on the specificity (TN rate = 0.52).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“However, ROC AUC presents several flaws [92], and it is sensitive to class imbalance [93].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [92] Lobo JM, Jiménez-Valverde A, Real R. AUC: a misleading measure of the performance of predictive distribution models. Glob Ecol Biogeogr. 2008; 17(2):145–51. source
- Not checked yetcites [93] Hanczar B, Hua J, Sima C, Weinstein J, Bittner M, Dougherty ER. Small-sample precision of ROC-related estimates. Bioinformatics. 2010; 26(6):822–30. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Once again, this score would hide the truth: the classification algorithm has performed poorly on the negative subset.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one. The quote only partly states this claim.
None of the sources cited for this claim have been checked yet.
In the paper
“The effectiveness of MCC has been shown in other scientific fields as well [43, 44].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [43] Liu Y, Cheng J, Yan C, Wu X, Chen F. Research on the Matthews correlation coefficients metrics of personalized recommendation algorithm evaluation. Int J Hybrid Informa Technol. 2015; 8(1):163–72. source
- Not checked yetcites [44] Naulaerts S, Dang CC, Ballester PJ. Precision and recall oncology: combining multiple gene mutations for improved identification of drug-sensitive tumours. Oncotarget. 2017; 8(57):97025. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“Even if less common than the ROC curve, several scientists consider the PR curve more informative than the ROC curve, especially on imbalanced biological and medical datasets [48, 97, 98].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [48] Saito T, Rehmsmeier M. The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE. 2015; 10(3):0118432. source
- Not checked yetcites [97] Chicco D. Ten quick tips for machine learning in computational biology. BioData Min. 2017; 10(35):1–17.
- Not checked yetcites [98] Ozenne B, Subtil F, Maucort-Boulch D. The precision–recall curve overcame the optimism of the receiver operating characteristic curve in rare diseases. J Clin Epidemiol. 2015; 68(8):855–9. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“MCC was re-proposed by Baldi and colleagues [39] in 2000 as a standard performance metric for machine learning with a natural extension to the multiclass case [40].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [39] Baldi P, Brunak S, Chauvin Y, Andersen CA, Nielsen H. Assessing the accuracy of prediction algorithms for classification: an overview. Bioinformatics. 2000; 16(5):412–24. source
- Not checked yetcites [40] Gorodkin J. Comparing two K-category assignments by a K-category correlation coefficient. Comput Biol Chem. 2004; 28(5–6):367–74. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“For example, when Whalen and colleagues released TargetFinder, a tool to predict enhancer-promoters interactions in genomics, they showed its results measured only by F1 score [13], making it impossible to detect the actual true positive rate and true negative rate of their tests [69].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [13] Whalen S, Truty RM, Pollard KS. Enhancer–promoter interactions are encoded by complex genomic signatures on looping chromatin. Nat Genet. 2016; 48(5):488. source
- Not checked yetcites [69] Xi W, Beer MA. Local epigenomic state cannot discriminate interacting and non-interacting enhancer–promoter pairs with high accuracy. PLoS Comput Biol. 2018; 14(12):1006625. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Although being widely acknowledged as a reliable metric, there are situations - albeit extreme - where either MCC cannot be defined or it displays large fluctuations [45], due to imbalanced outcomes in the classification.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [45] Brown JB. Classifiers and their metrics quantified. Mol Inform. 2018; 37:1700127. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“An effective solution overcoming the class imbalance issue comes from the Matthews correlation coefficient (MCC), a special case of the ϕ phi coefficient [36].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [36] Guilford JP. Psychometric Methods. New York City: McGraw-Hill; 1954.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“for instance, the Food and Drug Administration (FDA) agency of the USA employed the MCC as the main evaluation measure in the MicroArray II / Sequencing Quality Control (MAQC/SEQC) projects [41, 42].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [41] The MicroArray Quality Control (MAQC) Consortium. The MAQC-II Project: a comprehensive study of common practices for the development and validation of microarray-based predictive models. Nat Biotechnol. 2010; 28(8):827–38. source
- Not checked yetcites [42] The SEQC/MAQC-III Consortium. A comprehensive assessment of RNA-seq accuracy, reproducibility and information content by the Sequence Quality Control consortium. Nat Biotechnol. 2014; 32:903–14. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“However, this procedure is biased [65], and it is still far from being accepted as a standard practice by the community.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [65] Van Asch V. Macro-and micro-averaged evaluation measures. Technical report. 2013:1–27.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In this case, the algorithm showed its ability to predict the positive data instances (90 sick patients out of 91 were correctly predicted), but it also displayed its lack of talent in identifying healthy controls (only 1 healthy individual out of 9 was correctly recognized) (Fig. 2b). Therefore, the overall performance should be judged poor.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plaindirectSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Similar to the Use case A1 and B1, the method was able to correctly identify multiple negative data instances (88 healthy patients out of 89), but unable to correctly predict most of sick patients (only 2 true positives out of 11 possible elements).”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plainn = 88- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Originally developed by Matthews in 1975 for comparison of chemical structures [38]”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [38] Matthews BW. Comparison of the predicted and observed secondary structure of T4 phage lysozyme. Biochim Biophys Acta (BBA) Protein Struct. 1975; 405(2):442–51. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“Despite its popularity, in the learning context there are a number of issues causing the kappa measure to produce unreliable results (for instance, its high sensitivity to the distribution of the marginal totals [73–75])”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Background
What the sources say
- Not checked yetcites [73] Flight L, Julious SA. The disagreeable behaviour of the kappa statistic. Pharm Stat. 2015; 14:74–8. source
- Not checked yetcites [74] Powers DMW. The problem with Kappa. In: Proceedings of EACL 2012 – the 13th Conference of the European Chapter of the Association for Computational Linguistics. Avignon: ACL: 2012. p. 345–55.
- Not checked yetcites [75] Delgado R, Tibau X-A. Why Cohen’s Kappa should be avoided as performance measure in classification. PloS ONE. 2019; 14(9):0222916.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In fact, when a prediction displays many true positives but few true negatives (or many true negatives but few true positives) we will show that F1 and accuracy can provide misleading information, while MCC always generates results that reflect the overall prediction issues.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“These results show that, while accuracy and F1 score often generate high scores that do not inform the user about ongoing prediction issues, the MCC is a robust, useful, reliable, truthful statistical measure able to correctly reflect the deficiency of any prediction in any dataset.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
boosteddirectn = 6- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one. The quote only partly states this claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Here accuracy again generated overoptimistic, misleading, and inflated high results”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plaindirectSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The F1 score ranking and the accuracy ranking, in conclusion, are hiding this important flaw of the top classifier: k-NN was unable top correctly predict a high percentage of patients.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“There, accuracy and F1 generated overoptimistic and inflated results”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Accuracy and F1 score, although popular, can generate misleading results on imbalanced datasets, because they fail to consider the ratio between positive and negative elements.”
- Also stated in: § Conclusions · plain
“F1 and accuracy, instead, generate reliable results only when applied to balanced datasets, and produce misleading results when applied to imbalanced cases.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Conclusions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Also stated in: § Conclusions · plain
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“With its low value (MCC = +0.29), the MCC informs the readers about the poor general outcome of the radial SVM, while the accuracy and F1 score show misleading values.”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“The dataset contains 2,000 gene probsets for 62 patients, of which 22 are healthy controls and 40 have colon cancer (35.48% negatives and 64.52% positives) [106].”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
What the sources say
- Not checked yetcites [106] Alon U, Barkai N, Notterman DA, Gish K, Ybarra S, Mack D, Levine AJ. Data pertaining to the article ‘Broad patterns of gene expression revealed by clustering of tumor and normal colon tissues probed by oligonucleotide arrays’. 2000. http://genomics-pubs.princeton.edu/oncology/affydata/index.html. Accessed 10 Dec 2019.
plainn = 62- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper relies on this without support.
Accuracy and F1 score are the most common metrics used for binary classification in machine learning.
“We decided to focus on accuracy and F1 score because they are the most common metrics used for binary classification in machine learning.”
The paper relies on this without support.
The synthetic use cases are realistic representations of binary classification scenarios.
“here we describe some synthetic, realistic scenarios where MCC results are more informative and truthful than the other two measures analyzed.”
The paper relies on this without support.
Accuracy and F1 score are more commonly used in machine learning studies related to biomedical applications than other rates.
“because accuracy and F1 score are more commonly used in machine learning studies related to biomedical applications.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“MCC is a more reliable metric, whereby a high score is assigned only if the prediction was correct for the four base categories (TP, FP, TN and FP), taking into account the size of P and N [85].”
From A Survey of Underwater Acoustic Data Classification Methods Using Deep Learning for Shoreline Surveillance (2022) · cites this paper as [85]
In the citing paper
“since it has been shown to present a balanced score when dealing with imbalanced data [11, 6] as is our case with fewer manipulated images.”
From BioFors: A Large Biomedical Image Forensics Dataset (2021) · cites this paper as [11]
In the citing paper
“(3) the absolute value of F1 changes by simply switching the outcome labels (ie, when 1 becomes 0 and 0 becomes 1).”
From Evaluation of performance measures in predictive artificial intelligence models to support medical decisions: overview and guidance (2025) · cites this paper as [77]
In the citing paper
“(2) F1 has no intuitive interpretation”
From Evaluation of performance measures in predictive artificial intelligence models to support medical decisions: overview and guidance (2025) · cites this paper as [77]
In the citing paper
“(1) F1 ignores true negatives”
From Evaluation of performance measures in predictive artificial intelligence models to support medical decisions: overview and guidance (2025) · cites this paper as [77]
In the citing paper
“F1 and MCC were introduced to address the challenges caused by class imbalance, where classification accuracy can be inflated by classifying all patients as low risk when the event is rare.”
From Evaluation of performance measures in predictive artificial intelligence models to support medical decisions: overview and guidance (2025) · cites this paper as [77]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“Here, accuracy again generated an overoptimistic inflated score”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“all three statistical scores generated low results that reflect the mediocre quality of the prediction: F1 score = 0.17 and accuracy = 0.10 in the [0, 1] interval”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“the prediction technique failed in predicting negative elements, and therefore its confusion matrix showed a low percentage of true negatives”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“the method was able to correctly identify multiple negative data instances (88 healthy patients out of 89), but unable to correctly predict most of sick patients (only 2 true positives out of 11 possible elements)”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“Even if accuracy showed an excessively high result score”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“all three statistical rates showed a low score which emphasized the deficiency in the prediction process (accuracy = 0.24, F1 score = 0.12, and MCC = −0.24)”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
“the best classifier resulted being k-NN (F1 score = 0.87 and accuracy = 0.81)”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Genomics scenario: colon cancer gene expression
“k-nearest neighbors | +0.48 | 0.87 | 0.80 | 0.92 | 0.52”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“only 1 healthy individual out of 9 was correctly recognized”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“TP=90, FN=1, TN=0, FP=9”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“Finally, in the last Use case C2 (positively imbalanced dataset)”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“Use case C2 Negatively imbalanced dataset | 11 | 89 | 2 | 9 | 88 | 1”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“For the Use case A1 (negatively imbalanced dataset)”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“made of 9 healthy individuals (negatives =9%) and 91 sick patients (positives =91%)”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“Also in the Use case C1 (positively imbalanced dataset)”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“90 healthy controls (negatives =90%) and 10 sick patients (positives =10%)”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“only 5 sick patients out of 70”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“TP = 5, FN = 70, TN = 19, FP = 6”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Results and discussion > Use cases
“the resulting MCC = –0.03”
§ The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation > Methods > Relationship between measures
“In red, the (−0.04, 0.95) point corresponding to use case A1”
Citations
Cites
Breiman, L. (2001) · Machine Learning · doi:10.1023/a:1010933404324 · Cited by 133,538
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
The Measurement of Observer Agreement for Categorical Data
Landis, J. R., & Koch, G. G. (1977) · Biometrics · doi:10.2307/2529310 · Cited by 81,148
XGBoost: A Scalable Tree Boosting System
Chen, T., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939785 · Cited by 54,569
A Coefficient of Agreement for Nominal Scales
Cohen, J. (1960) · Educational and Psychological Measurement · doi:10.1177/001316446002000104 · Cited by 42,398
An introduction to ROC analysis
Fawcett, T. (2006) · Pattern Recognition Letters · doi:10.1016/j.patrec.2005.10.010 · Cited by 21,879
The meaning and use of the area under a receiver operating characteristic (ROC) curve.
Hanley, J. A., & McNeil, B. J. (1982) · Radiology · doi:10.1148/radiology.143.1.7063747 · Cited by 21,856
Interrater reliability: The kappa statistic
McHugh, M. L. (2012) · Biochemia Medica · doi:10.11613/bm.2012.031 · Cited by 19,877
Measures of the Amount of Ecologic Association Between Species
Dice, L. R. (1945) · Ecology · doi:10.2307/1932409 · Cited by 12,187
Statistical Comparisons of Classifiers over Multiple Data Sets
Demšar, J. (2006) · doi:10.5555/1248547.1248548 · Cited by 12,051
Research works printed but not linked: 7
- Akosa JS. Predictive accuracy: a misleading performance measure for highly imbalanced data. In: Proceedings of the SAS Global Forum 2017 Conference. Cary, North Carolina: SAS Institute Inc.: 2017. p. 942–2017.
- Guilford JP. Psychometric Methods. New York City: McGraw-Hill; 1954.
- van Rijsbergen CJ, Joost C. Information Retrieval. New York City: Butterworths; 1979.
- Sasaki Y. The truth of the F-measure. Teach Tutor Mater. 2007; 1(5):1–5.
- Dubey A, Tarar S. Evaluation of approximate rank-order clustering using Matthews correlation coefficient. Int J Eng Adv Technol. 2018; 8(2):106–13.
- Suresh Babu N. Various performance measures in binary classification – An overview of ROC study. Int J Innov Sci Eng Technol. 2015; 2(9):596–605.
- Ferri C, Hernández-Orallo J, Flach PA. A coherent interpretation of AUC as a measure of aggregated classification performance. In: Proceedings of ICML 2011 – the 28th International Conference on Machine Learning. Norristown: Omnipress: 2011. p. 657–64.
Cited by
Sorted by citations
Items 901–1,000 of 1,016
Dos and Don'ts of Machine Learning in Computer SecurityPreprint
Arp, D., Quiring, E., Pendlebury, F., Warnecke, A., Pierazzi, F., Wressnegger, C., Cavallaro, L., & Rieck, K. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.09470 · Cited by 222
“In these cases, other measures like Matthews Correlation Coefficient (MCC) are more suitable to assess the classifier’s performance [29] (see §4).”
Krishnan, R., Howard, I. S., Comber, S. D. W., & Jha, A. N. (2023) · The Science of The Total Environment · doi:10.1016/j.scitotenv.2023.164072 · Cited by 24
Financial sentiment analysis: Classic methods vs. deep learning models
Karanikola, A., Davrazos, G., Liapis, C. M., & Kotsiantis, S. B. (2023) · Intelligent Decision Technologies · doi:10.3233/idt-230478 · Cited by 30
“MCC is considered a superior metric to F1 score and accuracy, particularly in datasets with imbalanced classes [79].”
Methodology for biomarker discovery with reproducibility in microbiome data using machine learning
Rojas-Velázquez, D., Kidwai, S., Kraneveld, A. D., Tonda, A. P., Oberski, D. L., Garssen, J., & Lopez‐Rincon, A. (2024) · BMC Bioinformatics · doi:10.1186/s12859-024-05639-3 · Cited by 26
“Additionally, we use the Matthews Correlation Coefficient (MCC) [64] as a metric to evaluate the performance of the methodology as well as for comparison with other feature selection methods.”
MethodLUNet: deep learning for the segmentation of arterioles and venules in high resolution fundus images
Fhima, J., Eijgen, J. V., Moulin-Romsée, M.-I. B., et al. (2024) · Physiological Measurement · doi:10.1088/1361-6579/ad3d28 · Cited by 23
Bastien, G. E., Cable, R. N., Zaman, L., Batterbee, C., Wing, A. J., & Duhaime, M. B. (2024) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1011649 · Cited by 24
SPOT: A machine learning model that predicts specific substrates for transport proteins
Kroll, A., Niebuhr, N., Butler, G., & Lercher, M. J. (2024) · PLoS Biology · doi:10.1371/journal.pbio.3002807 · Cited by 29
Qadir, M. M. F., Elgamal, R. M., Song, K., Kudtarkar, P., Sakamuri, S. S. V. P., Katakam, P. V. G., El-Dahr, S. S., Kolls, J. K., Gaulton, K. J., & Mauvais‐Jarvis, F. (2024) · The EMBO Journal · doi:10.1038/s44318-024-00313-z · Cited by 25
“Infact, pseduobulk DEG methods demonstrate the highest Mathews Correlation Coef fi cient, a balanced machine learning performance testing model, capable of evaluating models classifying binary data (Chicco and Jurman, 2020; Murphy and Skene, 2022).”
Fair and Transparent Student Admission Prediction Using Machine Learning Models
Raftopoulos, G., Davrazos, G., & Kotsiantis, S. B. (2024) · Algorithms · doi:10.3390/a17120572 · Cited by 32
“For this reason, MCC is often recommended over simpler metrics for scenarios where class imbalance is significant, as it provides a clearer picture of a model’s capability to generalize across both majority and minority classes [40].”
Innab, N., Alsubai, S., Alabdulqader, E. A., Alarfaj, A. A., Umer, M., Treľová, S., & Ashraf, I. (2024) · Frontiers in Public Health · doi:10.3389/fpubh.2024.1462693 · Cited by 29
Evaluating machine learning models for stroke prediction based on clinical variables
Akinwumi, P. O., Ojo, S., Nathaniel, T. I., Wanliss, J. A., Karunwi, O., & Sulaiman, M. (2025) · Frontiers in Neurology · doi:10.3389/fneur.2025.1668420 · Cited by 31
“…prediction among all positive predictions) is shown in Equation 6: Recall (sensitivity), the model’s ability to identify actual stroke cases, crucial in medical diagnosis is shown in Equation 7: This measure is especially useful when false negatives and false positives carry similar costs (156).”
Suda, E. Y., Watari, R., Matias, A. B., et al. (2020) · Frontiers in Bioengineering and Biotechnology · doi:10.3389/fbioe.2020.00576 · Cited by 24
“Nevertheless, this issue was dealt by applying the MCC as the main performance measure because it is more reliable and informative when evaluating binary classifications, especially on imbalanced datasets (Chicco and Jurman, 2020).”
MethodA wheeze recognition algorithm for practical implementation in children
Habukawa, C., Ohgami, N., Matsumoto, N., Hashino, K., Asai, K., Sato, T., & Murakami, K. (2020) · PLoS ONE · doi:10.1371/journal.pone.0240048 · Cited by 24
“We analyzed the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) using the results of the wheeze recognition algorithm in all data files [26, 27].”
MethodChen, K., Vantuch, T., Zhang, Y., et al. (2021) · eScholarship (California Digital Library) · Cited by 23
Hybrid Elman Neural Network and an Invasive Weed Optimization Method for Bridge Defect Recognition
Abdelkader, E. M., Moselhi, O., Marzouk, M. M., & Zayed, T. M. (2020) · Transportation Research Record Journal of the Transportation Research Board · doi:10.1177/0361198120967943 · Cited by 22
Privacy Protection of Grid Users Data with Blockchain and Adversarial Machine Learning
Yilmaz, I., Kapoor, K., Siraj, A., et al. (2021) · doi:10.1145/3445969.3450431 · Cited by 23
“MCC ensures reliable assessment when a dataset is highly imbalanced such as in this case [24], [25], [26].”
Landslide detection in real-time social media image streams
Ofli, F., Imran, M., Qazi, U., Roch, J., Pennington, C., Banks, V., & Bossu, R. (2023) · Neural Computing and Applications · doi:10.1007/s00521-023-08648-0 · Cited by 29
Ko, A. L., Tong, A. P., Mossa‐Basha, M., Weaver, K. E., Ojemann, J. G., Miller, J. W., & Hakimian, S. (2021) · Epilepsia · doi:10.1111/epi.17059 · Cited by 23
Meniscal Tear and ACL Injury Detection Model Based on AlexNet and Iterative ReliefF
Key, S., Baygin, M., Demir, S., et al. (2022) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-022-00581-3 · Cited by 24
Park, B., Shin, T.-S., Cho, J., et al. (2021) · Agronomy · doi:10.3390/agronomy12010085 · Cited by 23
Ensembling perturbation-based oversamplers for imbalanced datasets
Zhang, J., Wang, T., Ng, W. W. Y., & Witold Pedrycz, C. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.01.049 · Cited by 24
Distributed radar fusion and recurrent networks for classification of continuous human activities
Guendel, R. G., Fioranelli, F., & Yarovoy, A. G. (2022) · IET Radar Sonar & Navigation · doi:10.1049/rsn2.12249 · Cited by 23
Tanko, D., Barua, P. D., Doğan, Ş., Tuncer, T., Palmer, E. E., Ciaccio, E. J., & Acharya, U. R. (2022) · Physiological Measurement · doi:10.1088/1361-6579/ac59dc · Cited by 24
Unsupervised Method for Wildfire Flame Segmentation and Detection
Buza, E., & Akagić, A. (2022) · IEEE Access · doi:10.1109/access.2022.3176724 · Cited by 23
Cao, J., You, K., Zhou, J., et al. (2022) · EClinicalMedicine · doi:10.1016/j.eclinm.2022.101633 · Cited by 23
Leveraging deep feature learning for wearable sensors based handwritten character recognition
Singh, S. K., & Chaturvedi, A. (2022) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2022.104198 · Cited by 23
Ågren, A. M., Hasselquist, E. M., Stendahl, J., Nilsson, M. B., & Paul, S. S. (2022) · SOIL · doi:10.5194/soil-8-733-2022 · Cited by 24
A Comparative Approach to Threshold Optimization for Classifying Imbalanced Data
Hancock, J., Johnson, J. M., & Khoshgoftaar, T. M. (2022) · IEEE International Conference on Collaboration and Internet Computing (CIC) · doi:10.1109/cic56439.2022.00028 · Cited by 24
“In this study, the function parameter may be one of the four classification metrics: geometric mean, MCC, precision, or f-measure.”
BackgroundInfluentialAmanullah, M. A., Loke, S. W., Chhetri, M. B., & Doss, R. (2023) · ACM Computing Surveys · doi:10.1145/3596598 · Cited by 23
Lobo, F., González, M. S., Boto, A., & de la Lastra, J. M. P. (2023) · International Journal of Molecular Sciences · doi:10.3390/ijms241210270 · Cited by 26
“methods trained on the same dataset [56].”
Casiano, A. S., Mehta, D., Danciu, O. C., Patel, H., Banks, L., Zaidi, A., Buckley, J., Rauscher, G. H., Schulte, L., Weller, L., et al. (2023) · Scientific Reports · doi:10.1038/s41598-023-39215-1 · Cited by 24
TPpred-LE: therapeutic peptide function prediction based on label embedding
Lv, H., Yan, K., & Liu, B. (2023) · BMC Biology · doi:10.1186/s12915-023-01740-w · Cited by 23
“…metrics to evaluate the performance for each therapeutic peptide function prediction task in a one-versus-all form, including the area under the ROC curve (AUC) [17], Matthews’s correlation coefficient (MCC) [18], the F1 measure [19], and the K-category correlation coefficient (RkCC) [20].”
BackgroundMethodFrom Proteins to Ligands: Decoding Deep Learning Methods for Binding Affinity Prediction
Gorantla, R., Kubincová, A., Weiße, A. Y., & Mey, A. S. J. S. (2024) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.3c01208 · Cited by 24
“To evaluate the performance of contact map prediction methods, we used the F1 score, Matthews’ correlation coefficient (MCC), 47 and precision metrics.”
Yan, F., Huang, H., Pedrycz, W., & Hirota, K. (2023) · Applied Soft Computing · doi:10.1016/j.asoc.2023.111123 · Cited by 23
Conspiracy thinking and social media use are associated with ability to detect deepfakes
Nas, E., & de Kleijn, R. (2023) · Telematics and Informatics · doi:10.1016/j.tele.2023.102093 · Cited by 28
Awe, O. O., Opateye, G. O., Johnson, C. A. G., Tayo, O. T., & Dias, R. (2023) · STEAM-H: Science, Technology, Engineering, Agriculture, Mathematics & Health · doi:10.1007/978-3-031-41352-0_18 · Cited by 25
Misbehavior detection in intelligent transportation systems based on federated learning
Campos, E. M., Hernández-Ramos, J. L., González-Vidal, A., Baldini, G., & Skármeta, A. (2024) · Internet of Things · doi:10.1016/j.iot.2024.101127 · Cited by 30
Krishna, T. B., & Kokil, P. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3369155 · Cited by 26
“…method on fetal US images, most frequently employed performance evaluation metrics in medical image analysis field [39], such as accuracy (Acc) [39], precision (Pre), sensitivity (Sen) [39], specificity (Spec) [39], F1-score (F1), and Matthews correlation coefficient (MCC) [40], are utilized.”
MethodTran, T.-O., & Le, N. Q. K. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.108408 · Cited by 24
Latifi, M., Zali, R. B., Javadi, A. A., & Farmani, R. (2024) · Journal of Water Resources Planning and Management · doi:10.1061/jwrmd5.wreng-6334 · Cited by 24
Benchmarking Safety Monitors for Image Classifiers with Machine Learning
Ferreira, R., Arlat, J., Guiochet, J., & Waeselynck, H. (2021) · IEEE Pacific Rim International Symposium on Dependable Computing (PRDC) · doi:10.1109/prdc53464.2021.00012 · Cited by 25
Chemoinformatics and Machine Learning Approaches for Identifying Antiviral Compounds
John, L., Soujanya, Y., Mahanta, H. J., & Sastry, G. N. (2021) · Molecular Informatics · doi:10.1002/minf.202100190 · Cited by 23
MalBERT: Using Transformers for Cybersecurity and Malicious Software DetectionPreprint
Rahali, A., & Akhloufi, M. A. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.03806 · Cited by 57
“We chose MCC over F1-score for binary classification as recommended in this study [36].”
MethodPrediction of ROP Zones Using Deep Learning Algorithms and Voting Classifier Technique
Salih, N., Ksantini, M., Hussein, N., et al. (2023) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-023-00268-9 · Cited by 22
“Pretrained models such as VGG-19, ResNet-50, and Effi-cientNetB5 were trained using significant data sources like ImageNet, which has over a thousand categories and 1.2 million natural pictures [9].”
MethodInfluentialPreinterventional Third-Molar Assessment Using Robust Machine Learning
Carvalho, J. S., Lotz, M., Rubi, L., Unger, S., Pfister, T., Buhmann, J. M., & Stadlinger, B. (2023) · Journal of Dental Research · doi:10.1177/00220345231200786 · Cited by 22
“The accuracy, F1-score, precision, recall, and MCC values evaluated on the out-of-distribution data can be seen in Figure 3C.”
MethodInfluentialEken, A., Nassehi, F., & Eroğul, O. (2024) · Reviews in the Neurosciences · doi:10.1515/revneuro-2023-0117 · Cited by 25
MoodCapture: Depression Detection using In-the-Wild Smartphone Images
Nepal, S. K., Pillai, A., Wang, W., Griffin, T. Z., Collins, A. C., Heinz, M. V., Lekkas, D., Mirjafari, S., Nemesure, M. D., Price, G. D., et al. (2024) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642680 · Cited by 31
“In fact, MCC is preferred over F1 score in many binary classification problems [13]. where 𝑇𝑃,𝑇𝑁, 𝐹𝑃, and 𝐹𝑁 are true positives, true negatives, false positives, and false negatives, respectively.”
BackgroundPerski, O., Kale, D., Leppin, C., Okpako, T., Simons, D., Goldstein, S. P., Hekler, E. B., & Brown, J. (2024) · PLOS Digital Health · doi:10.1371/journal.pdig.0000594 · Cited by 25
“Future research should consider using the Matthews Correlation Coefficient for imbalanced datasets [72], in addition to devising appropriate cut-offs for use in the context of a smoking lapse estimator, which are currently lacking.”
Enhanced prediction of pipe failure through transient simulation-aided logistic regression
Zhong, D., Huang, C., Ma, W., et al. (2025) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2025.110913 · Cited by 22
AD-Diff: enhancing Alzheimer's disease prediction accuracy through multimodal fusion
Han, L. (2025) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2025.1484540 · Cited by 28
“Its removal caused a comprehensive decline in performance, especially in terms of precision and MCC metrics.”
InfluentialRetractedTrust and Transparency in Artificial Intelligence
Mitchell, T. J. (2025) · Philosophy & Technology · doi:10.1007/s13347-025-00916-2 · Cited by 33
“Other suggested measures of reliability include balanced accuracy, the Matthews correlation coefficient, the F*-measure, and the K-measure, to name but a few (Chicco & Jurman, 2020; Christen et al., 2023; Diallo et al., 2024).”
An improved agglomerative hierarchical clustering anomaly detection method for scientific data
Shi, P., Zhao, Z., Zhong, H., Shen, H., & Ding, L. (2020) · Concurrency and Computation Practice and Experience · doi:10.1002/cpe.6077 · Cited by 22
Zincbindpredict—Prediction of Zinc Binding Sites in Proteins
Ireland, S. M., & Martin, A. C. R. (2021) · Molecules · doi:10.3390/molecules26040966 · Cited by 22
“For the trained model, the metrics recall (how effective at finding true binding sites the models are), precision (how effective at ignoring non-binding sites they are), the F1 score (the harmonic mean of recall and precision) and Matthews Correlation Coefficient (another summary of the true positives, true negatives,…”
MethodA Deep Learning-Based Recognition Approach for the Conversion of Multilingual Braille Images
Al‐Salman, A. S., Gumaei, A., Al‐Salman, A. S., et al. (2021) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2021.015614 · Cited by 22
Legal document recommendation system: A cluster based pairwise similarity computation
Dhanani, J., Mehta, R. G., & Rana, D. P. (2021) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-189871 · Cited by 22
“This research evaluates the performance of the proposed approach using binary classifica-tionmeasureslikeAccuracy,F1-Score,andMatthews CorrelationCoefficient(MCC)[17],arecomputed fromtheconfusionmatrixshowninFig.2.”
InfluentialProduct progression: a machine learning approach to forecasting industrial upgrading
Albora, G., Pietronero, L., Tacchella, A., & Zaccaria, A. (2023) · Scientific Reports · doi:10.1038/s41598-023-28179-x · Cited by 25
“• Matthews coefficient Matthews’ correlation coefficient 73 takes in to account all the four classes of the confusion matrix and the class imbalance issue 74,75 .”
Backgroundde‐Marcos, L., Martínez-Herráiz, J.-J., Junquera-Sánchez, J., Cilleruelo, C., & Pages-Arévalo, C. (2021) · Electronics · doi:10.3390/electronics10141622 · Cited by 22
“MCC is a more reliable statistical metric that produces a high score only if both classes are predicted well, even if one class is disproportionately overrepresented [31].”
BackgroundAn Improved Method for Training Data Selection for Cross-Project Defect Prediction
Bhat, N. A., & Farooq, S. U. (2021) · Arabian Journal for Science and Engineering · doi:10.1007/s13369-021-06088-3 · Cited by 22
Akinnuwesi, B. A., Fashoto, S. G., Mbunge, E., Odumabo, A. T., Metfula, A. S., Mashwama, P. M., Uzoka, F., Owolabi, O., Okpeku, M., & Amusa, O. O. (2021) · Data Science and Management · doi:10.1016/j.dsm.2021.12.001 · Cited by 22
Incipient Residual-Based Anomaly Detection in Power Electronic Devices
Yang, Q., Gultekin, M. A., Seferian, V., Pattipati, K. R., Bazzi, A. M., Palmieri, F., Rajamani, R., Joshi, S. N., Farooq, M., & Ukegawa, H. (2022) · IEEE Transactions on Power Electronics · doi:10.1109/tpel.2022.3140721 · Cited by 23
“The performance of the approaches are evaluated using accuracy (33), precision (34), recall (35), F-measure (36) [70], and the Cohen’s Kappa statistic (37) [71] in Table VI. True and false positives ( TP / FP ) are the number of predicted positives that were correct/incorrect, and true and false…”
MethodKumar, A., & Gorai, A. K. (2022) · Geocarto International · doi:10.1080/10106049.2022.2057595 · Cited by 24
“The accuracy assessment of the classi fi cation model was carried out based on the four indices, including Precision, Recall, F1 score and Matthews ’ s correlation coef fi cient (MCC) (Junker et al. 1999; Chicco and Jurman 2020).”
MethodMachine Learning Improves Upon Clinicians' Prediction of End Stage Kidney Disease
Chuah, A., Walters, G. D., Christiadi, D., Karpe, K. M., Kennard, A. L., Singer, R. F., Talaulikar, G. S., Ge, W., Suominen, H., Andrews, T. D., et al. (2022) · Frontiers in Medicine · doi:10.3389/fmed.2022.837232 · Cited by 22
“Commonly, F1 score (the harmonic mean of precision and recall) has been used to provide a fairer measure of predictor performance, but recently the Matthews correlation coefficient has been shown (28) to be better than both F1 and accuracy as a binary classification metric.”
MethodLouk, M. H. L., & Tama, B. A. (2022) · Big Data and Cognitive Computing · doi:10.3390/bdcc6020041 · Cited by 21
Saha, T. K., Pal, S., & Sarda, R. (2022) · Environmental Science and Pollution Research · doi:10.1007/s11356-022-21072-6 · Cited by 22
“MCC was used to check the quality of binary classification of machine learning models (Chicco and Jurman 2020).”
RetractedPrasad, B., McGeough, C. M., Eakin, A., Ahmed, T., Small, D., Gardiner, P. V., Pendleton, A., Wright, G. D., Bjourson, A. J., Gibson, D. S., et al. (2022) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1010204 · Cited by 23
“Further, the mean Matthews correlation coefficient (MCC), popularly used and advocated to assess the quality of binary classification [54], was 0.”
MethodTowards using visual, semantic and structural features to improve code readability classification
Mi, Q., Hao, Y., Ou, L., & Ma, W. (2022) · Journal of Systems and Software · doi:10.1016/j.jss.2022.111454 · Cited by 26
Impact of scanner variability on lymph node segmentation in computational pathology
Khan, A., Janowczyk, A. R., Müller, F. C., Blank, A., Nguyen, H. G., Abbet, C., Studer, L., Lugli, A., Dawson, H., Thiran, J., et al. (2022) · Journal of Pathology Informatics · doi:10.1016/j.jpi.2022.100127 · Cited by 23
“Pixel-level differences are measured using the Matthews correlation coefficient (MCC).(34,35) The MCC between ground truth and segmented labels are calculated as given in Eq.”
Target-Oriented Fusion of Attributes in Data Level for Salt Dome Geobody Delineation in Seismic Data
Khayer, K., Kahoo, A. R., Soleimani, M., et al. (2022) · Natural Resources Research · doi:10.1007/s11053-022-10086-z · Cited by 22
Coloma, S., Martínez, C., Yalçın, B. C., et al. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3198794 · Cited by 22
“This metric is more informative and truthful in evaluating binary classifications than accuracy or F1-score, especially when leading with imbalanced datasets [23].”
BackgroundMetabolomic and exposomic biomarkers of risk of future neurodevelopmental delay in human milk
Li, K., Bertrand, K. A., Naviaux, J. C., Monk, J. M., Wells, A., Wang, L., Lingampelly, S. S., Naviaux, R. K., & Chambers, C. (2022) · Pediatric Research · doi:10.1038/s41390-022-02283-6 · Cited by 22
“Matthews correlation coefficient (MCC) was also calculated, which provided complementary information to common binary classification metrics, especially for class-imbalanced datasets.(33) See details in the Supplementary Methods.”
MethodAutomated Atrial Fibrillation Detection with ECG
Wei, T.-R., Lu, S., & Yan, Y. (2022) · Bioengineering · doi:10.3390/bioengineering9100523 · Cited by 24
“Furthermore, we include the Matthews correlation coefficient (MCC) along with other metrics to evaluate our model performance, since MCC is considered the best metric to evaluate classifiers’ performances, especially on imbalanced datasets [25,26].”
BackgroundMethodChoi, Y., An, J., Ryu, S., et al. (2022) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph192013672 · Cited by 22
“F1-score calculated through the confusion matrix to confirm the best performing model among all experimental models is a harmonic average of precision and recall and is an evaluation index mainly used when data imbalance between classification classes is severe [32].”
MethodVerschoof, W. B., Bonhage, A., Schneider, A., Ouimet, W. B., & Raab, T. (2022) · Archaeological Prospection · doi:10.1002/arp.1889 · Cited by 23
Disentangled Representation for Cross-Domain Medical Image Segmentation
Wang, J., Zhong, C., Feng, C., Zhang, Y., Sun, J., & Yokota, Y. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3221131 · Cited by 22
“We also report the Matthews correlation coefficient (MCC) [26], [58] similar to [26], considering that MCC”
MethodPlexusNet: A neural network architectural concept for medical image classification
Eminağa, O., Abbas, M., Shen, J., Laurie, M. A., Brooks, J. D., Liao, J. C., & Rubin, D. L. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.106594 · Cited by 27
Casal-Guisande, M., Torres‐Durán, M., Mosteiro, M., Cerqueiro-Pequeño, J., Bouza-Rodríguez, J.-B., Fernández‐Villar, A., & Comesaña-Campos, A. (2023) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph20043627 · Cited by 22
“In this sense, the Matthews correlation coefficient value lies in the interval [−1 1], with −1 and 1 corresponding respectively to the perfect error and the perfect success in the classification, while the intermediate point 0 indicates a classification based on pure chance [47].”
BackgroundMethodWFT-Fati-Dec: Enhanced Fatigue Detection AI System Based on Wavelet Denoising and Fourier Transform
Sedik, A., Marey, M., & Mostafa, H. (2023) · Applied Sciences · doi:10.3390/app13052785 · Cited by 22
OWSum: algorithmic odor prediction and insight into structure-odor relationships
Schicker, D., Singh, S., Freiherr, J., & Grasskamp, A. T. (2023) · Journal of Cheminformatics · doi:10.1186/s13321-023-00722-y · Cited by 22
Dipaola, F., Gatti, M., Levra, A. G., Menè, R., Shiffer, D., Faccincani, R., Raouf, Z., Secchi, A., Rovere, P. Q., Voza, A., et al. (2023) · Scientific Reports · doi:10.1038/s41598-023-37512-3 · Cited by 22
“Cough, n (%) 672 (52) 290 (57) 382 (49)”
BTLink : automatic link recovery between issues and commits based on pre-trained BERT model
Lan, J., Gong, L., Zhang, J., & Zhang, H. (2023) · Empirical Software Engineering · doi:10.1007/s10664-023-10342-7 · Cited by 25
“In addition to Precision, Recall, F-Measure, and MCC (Matthews correlation coefficient) (Chicco and Jurman 2020) commonly used in previous studies, we also consider some measures commonly used in classification models, such as AUC (Area Under the ROC Curve), ACC (Accuracy), and PF (Probability of false alarm).”
BackgroundMethodInfluentialRecent Advances in Toxicity Prediction: Applications of Deep Graph Learning
Miao, Y., Ma, H., & Huang, J. (2023) · Chemical Research in Toxicology · doi:10.1021/acs.chemrestox.2c00384 · Cited by 29
Federated Learning in Computational Toxicology: An Industrial Perspective on the Effiris Hackathon
Bassani, D., Brigo, A., & Morger, A. (2023) · Chemical Research in Toxicology · doi:10.1021/acs.chemrestox.3c00137 · Cited by 25
Diniz, A. P. M., Ciarelli, P. M., Salles, E. O. T., & Côco, K. F. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.121963 · Cited by 24
KinomeMETA: meta-learning enhanced kinome-wide polypharmacology profiling
Ren, Q., Qu, N., Sun, J., Zhou, J., Liu, J., Ni, L., Tong, X., Zhang, Z., Kong, X., Wen, Y., et al. (2023) · Briefings in Bioinformatics · doi:10.1093/bib/bbad461 · Cited by 24
Azis, H., Nirmala, Syafie, L., et al. (2024) · International Conference on Ubiquitous Information Management and Communication (IMCOM) · doi:10.1109/imcom60618.2024.10418246 · Cited by 22
“Recall (Sensitivity or True Positive Rate) gauges the extent to which the model can correctly recognize and classify actual positive cases [45] by comparing the predicted positives to the actual positives.”
BackgroundA heart disease prognosis pipeline for the edge using federated learning
Mahalingam, P. R., & Dheeba, J. (2024) · e-Prime - Advances in Electrical Engineering Electronics and Energy · doi:10.1016/j.prime.2024.100490 · Cited by 23
Genome annotation across species using deep convolutional neural networks
Khodabandelou, G., Routhier, E., & Mozziconacci, J. (2020) · PeerJ Computer Science · doi:10.7717/peerj-cs.278 · Cited by 23
Gajewicz, A., Gromelski, M., Wyrzykowska, E., Furuhama, A., Yamamoto, H., & Suzuki, N. (2020) · Ecotoxicology and Environmental Safety · doi:10.1016/j.ecoenv.2020.111738 · Cited by 23
Decision boundary clustering for efficient local SVM
Fayed, H. A., & Atiya, A. F. (2021) · Applied Soft Computing · doi:10.1016/j.asoc.2021.107628 · Cited by 24
One-Class Anomaly Detection for Industrial Applications: A Comparative Survey and Experimental Study
Paolini, D., Dini, P., Soldaini, E., & Saponara, S. (2025) · Computers · doi:10.3390/computers14070281 · Cited by 21
Smajíć, A., Rami, I., Sosnin, S., & Ecker, G. F. (2023) · Chemical Research in Toxicology · doi:10.1021/acs.chemrestox.3c00042 · Cited by 22
Naglik, I., & Lango, M. (2023) · Machine Learning · doi:10.1007/s10994-023-06416-8 · Cited by 22
“F-score is very often used and recommended in some application areas (Baeza-Yates and Ribeiro-Neto, 1999), but some works argue that Matthews correlation coefficient is a more reliable classification performance indicator while working with imbalanced data (Chicco & Jurman, 2020).”
Integrating SAR, Optical, and Machine Learning for Enhanced Coastal Mangrove Monitoring in Guyana
Chan-Bagot, K., Herndon, K. E., Nicolau, A. P., et al. (2024) · Remote Sensing · doi:10.3390/rs16030542 · Cited by 21
“The F1-score is calculated using the precision and recall, also known as producer’s and user’s accuracy [65,66].”
MethodRoad Crash Injury Severity Prediction Using a Graph Neural Network Framework
Sattar, K. A., Ishak, I., Affendey, L. S., et al. (2024) · IEEE Access · doi:10.1109/access.2024.3373885 · Cited by 21
“It differed from accuracy in a way that it produced a high value only in cases where it could correctly predict the majority of positive records as well as negative records [78].”
BackgroundAhmad, I., Wan, Z., Ahmad, A., et al. (2024) · Mathematics · doi:10.3390/math12101437 · Cited by 21
“Another study used Q-learning to identify malware on mobile devices, and it achieved 67% accuracy [30].”
An AutoEncoder enhanced light gradient boosting machine method for credit card fraud detection
Ding, L., Liu, L., Wang, Y., et al. (2024) · PeerJ Computer Science · doi:10.7717/peerj-cs.2323 · Cited by 21
Fox, B., Jiang, J. X., Wickramaratne, S., Kovatch, P., Suárez‐Fariñas, M., Shah, N. A., Parekh, A., & Nadkarni, G. N. (2025) · SLEEP · doi:10.1093/sleep/zsaf061 · Cited by 32
Guan, J., Xie, P., Meng, D., Yao, L., Yu, D., Chiang, Y., Lee, T.-Y., & Wang, J. (2025) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2025.05.039 · Cited by 26
qian, T., Su, X., Wu, H., Singh, V. P., & Zhang, T. (2025) · Agricultural Water Management · doi:10.1016/j.agwat.2025.109582 · Cited by 21
Kaplan, E., Ekinci, T., Kaplan, S., Barua, P. D., Dogan, S., Tuncer, T., Tan, R.-S., Arunkumar, N., & Acharya, U. R. (2022) · Contrast Media & Molecular Imaging · doi:10.1155/2022/6034971 · Cited by 16
Metadata
- Authors
- Chicco, D., & Jurman, G.
- Affiliations
- Krembil Research Institute (CA)
- Fondazione Bruno Kessler (IT)
- Corresponding author
- Chicco, Davide
- Venue
- BMC Genomics
- Published
- 2020-01-02
- Volume
- 21
- Issue
- 1
- Pages
- 6-6
- ISSN
- 1471-2164 (ISSN-L)
- PMC
- PMC6941312
- OpenAlex
- W2999309192
- Semantic Scholar
- 44858b662822d291a5e1e9e3501ad19a9f1572c2
- Cited by
- 6,143
- References
- 104
- Citations by year
- 2026: 1118 · 2025: 1152 · 2024: 1060 · 2023: 1032 · 2022: 883 · 2021: 682 · 2020: 211
- FWCI
- 276.55
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- BMC Genomics (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- BOA (University of Milano-Bicocca) (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Iris (University of Trento) (submittedVersion)
- Funders
- University of Toronto
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Imbalanced Data Classification Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Keywords
- Matthews correlation coefficient, F1 score, accuracy, confusion matrix, class imbalance, classification metrics
- MeSH terms
- Machine Learning, Correlation of Data, Algorithms, Data Interpretation, Statistical, Computational Biology
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-09, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-09 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
