A semi-supervised approach using label propagation to support citation screening
Open accessThe DOI matches its Crossref record; 17 of the 17 references it cites were found in a registry or an index.
Abstract
Citation screening, an integral process within systematic reviews that identifies citations relevant to the underlying research question, is a time-consuming and resource-intensive task. During the screening task, analysts manually assign a label to each citation, to designate whether a citation is eligible for inclusion in the review. Recently, several studies have explored the use of active learning in text classification to reduce the human workload involved in the screening task. However, existing approaches require a significant amount of manually labelled citations for the text classification to achieve a robust performance. In this paper, we propose a semi-supervised method that identifies relevant citations as early as possible in the screening process by exploiting the pairwise similarities between labelled and unlabelled citations to improve the classification performance without additional manual labelling effort. Our approach is based on the hypothesis that similar citations share the same label (e.g., if one citation should be included, then other similar citations should be included also). To calculate the similarity between labelled and unlabelled citations we investigate two different feature spaces, namely a bag-of-words and a spectral embedding based on the bag-of-words. The semi-supervised method propagates the classification codes of manually labelled citations to neighbouring unlabelled citations in the feature space. The automatically labelled citations are combined with the manually labelled citations to form an augmented training set. For evaluation purposes, we apply our method to reviews from clinical and public health. The results show that our semi-supervised method with label propagation achieves statistically significant improvements over two state-of-the-art active learning approaches across both clinical and public health reviews.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Cortes, C., & Vapnik, V. (1995) · Machine Learning · doi:10.1007/bf00994018 · Cited by 59,421
PRISMA-P Group, Moher, D., Shamseer, L., Clarke, M., Ghersi, D., Liberati, A., Petticrew, M., Shekelle, P., & Stewart, L. A. (2015) · Systematic Reviews · doi:10.1186/2046-4053-4-1 · Cited by 26,223
A Sequential Algorithm for Training Text Classifiers
Lewis, D. D., & Gale, W. A. (1994) · SIGIR ’94 · doi:10.1007/978-1-4471-2099-5_1 · Cited by 1,208
Seventy-five trials and eleven systematic reviews a day: how will we ever keep up?
Bastian, H., Glasziou, P., & Chalmers, I. (2010) · PLoS Medicine · doi:10.1371/journal.pmed.1000326 · Cited by 1,188 · cited 2×
“Such reviews are becoming increasingly difficult to manually develop and update due to the exponential growth of the biomedical literature (e.g., on average 75 trials and 11 systematic reviews are published daily in MEDLINE [18]).”
BackgroundO’Mara-Eves, A., Thomas, J., McNaught, J., Miwa, M., & Ananiadou, S. (2015) · Systematic Reviews · doi:10.1186/2046-4053-4-5 · Cited by 681
Self-citationReducing Workload in Systematic Review Preparation Using Automated Citation Classification
Cohen, A. M., Hersh, W. R., Peterson, K., & Yen, P.-Y. (2006) · Journal of the American Medical Informatics Association · doi:10.1197/jamia.m1929 · Cited by 411 · cited 6×
“Such imbalanced datasets constitute challenging cases for machine learning methods [2,8,13,31].”
BackgroundMethodInfluentialSemi-automated screening of biomedical citations for systematic reviews
Wallace, B. C., Trikalinos, T. A., Lau, J., Brodley, C., & Schmid, C. H. (2010) · BMC Bioinformatics · doi:10.1186/1471-2105-11-55 · Cited by 351 · cited 15×
“Such imbalanced datasets constitute challenging cases for machine learning methods [2,8,13,31].”
BackgroundMethodInfluentialSWIFT-Review: a text-mining workbench for systematic review
Howard, B. E., Phillips, J., Miller, K., Tandon, A., Mav, D., Shah, M. R., Holmgren, S., Pelch, K. E., Walker, V., Rooney, A. A., et al. (2016) · Systematic Reviews · doi:10.1186/s13643-016-0263-z · Cited by 196 · cited 1×
“[28] developed a regularised log-linear classification model which exploits two types of features, namely bag-of-words weighted by TF-IDF and topic-based features extracted by Latent Dirichlet Allocation (LDA) [29].”
MethodReducing systematic review workload through certainty-based screening
Miwa, M., Thomas, J., O’Mara-Eves, A., & Ananiadou, S. (2014) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2014.06.005 · Cited by 190 · cited 10×
“The public health reviews were developed by the EPPI-Centre 2 and reused by Miwa et al. [8] to investigate the performance of both certainty and uncertainty-based active learners.”
BackgroundMethodInfluentialSelf-citationShemilt, I., Simon, A., Hollands, G. J., Marteau, T. M., Ogilvie, D., O'Mara‐Eves, A., Kelly, M. P., & Thomas, J. (2014) · Research Synthesis Methods · doi:10.1002/jrsm.1093 · Cited by 170
Estimating Time to Conduct a Meta-analysis From Number of Citations Retrieved
Allen, I. E. (1999) · JAMA: The Journal of the American Medical Association · doi:10.1001/jama.282.7.634 · Cited by 163 · cited 1×
“This amounts to a considerable human workload, given that a typical screening task involves manually screening thousands of citations [3, 4, 5].”
BackgroundThe automation of systematic reviews
Tsafnat, G., Dunn, A., Glasziou, P., & Coiera, E. (2013) · BMJ · doi:10.1136/bmj.f139 · Cited by 131
Automatic text classification to support systematic reviews in medicine
García Adeva, J. J., Pikatza Atxa, J. M., Ubeda Carrillo, M., & Ansuategi Zengotitabengoa, E. (2014) · Expert Systems with Applications · doi:10.1016/j.eswa.2013.08.047 · Cited by 125 · cited 1×
“[30] studied the contribution of different segments of a citation (e.”
BackgroundActive learning for biomedical citation screening
Wallace, B. C., Small, K., Brodley, C. E., & Trikalinos, T. A. (2010) · Proceedings of the 16th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/1835804.1835829 · Cited by 123 · cited 1×
“[43] noted that, according to experts, yield”
A new algorithm for reducing the workload of experts in performing systematic reviews
Matwin, S., Kouznetsov, A., Inkpen, D., Frunza, O., & O'Blenis, P. (2010) · Journal of the American Medical Informatics Association · doi:10.1136/jamia.2010.004325 · Cited by 100 · cited 1×
“Matwin et al. [26] explored the use of a factorised version of the na¨ıve Bayes classifier as opposed to the voted perceptron method used in [13].”
MethodScreening nonrandomized studies for medical systematic reviews: a comparative study of classifiers
Bekhuis, T., & Demner-Fushman, D. (2012) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2012.05.002 · Cited by 70 · cited 1×
“This observation was subsequently confirmed by Bekhuis and DemnerFushman [33].”
ResultAdvanced analytics for the automation of medical systematic reviews
Timsina, P., Liu, J., & El-Gayar, O. (2016) · Information Systems Frontiers · doi:10.1007/s10796-015-9589-7 · Cited by 33 · cited 1×
“[32] investigated different strategies to mitigate the class imbalance between eligible and ineligible citations which is known to affect the performance of the classification model.”
Background
Cited by
Automation of systematic literature reviews: A systematic literature review
van Dinter, R., Tekinerdogan, B., & Catal, C. (2021) · Information and Software Technology · doi:10.1016/j.infsof.2021.106589 · Cited by 432
Bannach-Brown, A., Przybyła, P., Thomas, J., Rice, A. S. C., Ananiadou, S., Liao, J., & Macleod, M. R. (2019) · Systematic Reviews · doi:10.1186/s13643-019-0942-7 · Cited by 185
“…if performance does not meet a required threshold then more records are screened, chosen either through random sampling or, using active learning [10], on the basis either of those with highest uncertainty of predictions [11, 12] or alternatively from those most likely to be included [13 – 15].”
BackgroundSelf-citationArtificial intelligence to automate the systematic review of scientific literature
de la Torre-López, J., Ramírez, A., & Romero, J. R. (2023) · Computing · doi:10.1007/s00607-023-01181-x · Cited by 170
“[33] use the label assigned to one paper by the human reviewer to tag other similar papers that remain unlabelled.”
BackgroundPrioritising references for systematic reviews with RobotAnalyst: A user study
Przybyła, P., Brockmeier, A. J., Kontonatsios, G., Le Pogam, M., McNaught, J., von Elm, E., Nolan, K., & Ananiadou, S. (2018) · Research Synthesis Methods · doi:10.1002/jrsm.1311 · Cited by 148
Self-citationA survey on clinical natural language processing in the United Kingdom from 2007 to 2022
Wu, H., Wang, M., Wu, J., Francis, F., Chang, Y.-H., Shavick, A., Dong, H., Poon, M. T. C., Fitzpatrick, N., Levine, A. P., et al. (2022) · npj Digital Medicine · doi:10.1038/s41746-022-00730-6 · Cited by 101
“…involved and industry involved project budgets are plotted alongside the budget of all projects across fi ve 3-year periods. review 57,114,117,119,120 , drug-related 58,100,116,118 , randomised clinical trials 127 – 129 and generic classi fi cations (such as classifying or clustering documents)…”
Background
Metadata
- Authors
- Kontonatsios, G., Brockmeier, A. J., Przybyła, P., McNaught, J., Mu, T., Goulermas, J. Y., & Ananiadou, S.
- Venue
- Journal of Biomedical Informatics
- Publisher
- Elsevier BV
- Published
- 2017-08
- Volume
- 72
- Pages
- 67-76
- ISSN
- 1532-0464 (ISSN-L)
- OpenAlex
- W2657315003
- Cited by
- 42
- References
- 17
- FWCI
- 4.33
- Type
- journal article
- Language
- en
- Open access
- hybrid
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.
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.