Deep learning and time series-to-image encoding for financial forecasting
Open accessThe DOI matches its Crossref record; 31 of the 31 references it cites were found in a registry or an index.
Abstract
In the last decade, market financial forecasting has attracted high interests amongst the researchers in pattern recognition. Usually, the data used for analysing the market, and then gamble on its future trend, are provided as time series; this aspect, along with the high fluctuation of this kind of data, cuts out the use of very efficient classification tools, very popular in the state of the art, like the well known convolutional neural networks (CNNs) models such as Inception, ResNet, AlexNet, and so on. This forces the researchers to train new tools from scratch. Such operations could be very time consuming. This paper exploits an ensemble of CNNs, trained over Gramian angular fields (GAF) images, generated from time series related to the Standard & Poor's 500 index future; the aim is the prediction of the future trend of the U.S. market. A multi-resolution imaging approach is used to feed each CNN, enabling the analysis of different time intervals for a single observation. A simple trading system based on the ensemble forecaster is used to evaluate the quality of the proposed approach. Our method outperforms the buyand-hold (B&H) strategy in a time frame where the latter provides excellent returns. Both quantitative and qualitative results are provided.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The text we found for this paper looks like another document, so the self-check wasn’t run.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“It consists of a simplified version of the VGG-16 network [31], composed by 5 convolutional layers and a fully-connected one.”
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 2×
“In a successive step, we applied the RGB color map to the image, thus resulting in a three channel matrix (further details on this process can be found in [28] and [30]).”
MethodCortes, C., & Vapnik, V. (1995) · Machine Learning · doi:10.1007/bf00994018 · Cited by 59,421
Cortes, C., & Vapnik, V. (1995) · Machine-mediated learning · doi:10.1023/a:1022627411411 · Cited by 48,445 · cited 1×
“In [16], the authors have compared the capabilities of the support vector machines [17] in market prediction related issues against those obtained by using the radial basis function (RBF) networks and back propagation (BP).”
BackgroundUsing artificial neural network models in stock market index prediction
Guresen, E., Kayakutlu, G., & Daim, T. U. (2011) · Expert Systems with Applications · doi:10.1016/j.eswa.2011.02.068 · Cited by 786 · cited 1×
“In [7], a similar approach has exploited eleven market indicators for building and training a recurrent neural network for monthly transition of the stock price index [8].”
BackgroundGao, S., Zhou, M., Wang, Y., Cheng, J., Yachi, H., & Wang, J. (2019) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2018.2846646 · Cited by 698 · cited 1×
“In [7], a similar approach has exploited eleven market indicators for building and training a recurrent neural network for monthly transition of the stock price index [8].”
BackgroundStock market prediction system with modular neural networks
Kimoto, T., Asakawa, K., Yoda, M., & Takeoka, M. (1990) · 1990 IJCNN International Joint Conference on Neural Networks · doi:10.1109/ijcnn.1990.137535 · Cited by 650
Evaluating multiple classifiers for stock price direction prediction
Ballings, M., Van den Poel, D., Hespeels, N., & Gryp, R. (2015) · Expert Systems with Applications · doi:10.1016/j.eswa.2015.05.013 · Cited by 611 · cited 1×
“A comparison of different market price prediction approaches is shown in [21], where classification-based approaches usually provide better results, such as [22]–[24], rather than some regression-based competitors.”
BackgroundPredicting stock market index using fusion of machine learning techniques
Patel, J., Shah, S., Thakkar, P., & Kotecha, K. (2015) · Expert Systems with Applications · doi:10.1016/j.eswa.2014.10.031 · Cited by 608 · cited 1×
“Interesting results have also been obtained when fusing together the above described techniques: as an example, in [19] the authors have fused ANN with decision trees (DT).”
MethodLiterature review: Machine learning techniques applied to financial market prediction
Henrique, B. M., Sobreiro, V. A., & Kimura, H. (2019) · Expert Systems with Applications · doi:10.1016/j.eswa.2019.01.012 · Cited by 558 · cited 1×
“In [16], the authors have compared the capabilities of the support vector machines [17] in market prediction related issues against those obtained by using the radial basis function (RBF) networks and back propagation (BP).”
BackgroundWang, Z., & Oates, T. (2015) · National Conference on Artificial Intelligence · Cited by 500
Imaging Time-Series to Improve Classification and Imputation
Wang, Z., & Oates, T. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1506.00327 · Cited by 470 · cited 1×
“Indeed, CNNs has shown very good accuracy results when applied to pattern recognition on image and video data [26], [27].”
MethodStock Market Analysis: A Review and Taxonomy of Prediction Techniques
Shah, D., Isah, H., & Zulkernine, F. (2019) · International Journal of Financial Studies · doi:10.3390/ijfs7020026 · Cited by 465 · cited 1×
“Similarly, in [20], a two-stage fusion approach is proposed for predicting CNX Nifty and S&P Bombay Stock Exchange (BSE) Sensex from Indian stock markets.”
BackgroundSentiment analysis of Twitter data for predicting stock market movements
Pagolu, V. S., Reddy, K. N., Panda, G., & Majhi, B. (2016) · 2016 International Conference on Signal Processing, Communication, Power and Embedded System (SCOPES) · doi:10.1109/scopes.2016.7955659 · Cited by 464
Stock market prediction of S&P 500 via combination of improved BCO approach and BP neural network
Zhang, Y., & Wu, L. (2009) · Expert Systems with Applications · doi:10.1016/j.eswa.2008.11.028 · Cited by 423
Oliveira, N., Cortez, P., & Areal, N. (2017) · Expert Systems with Applications · doi:10.1016/j.eswa.2016.12.036 · Cited by 398
Predicting the direction of stock market prices using tree-based classifiers
Basak, S., Kar, S., Saha, S., Khaidem, L., & Dey, S. R. (2019) · The North American Journal of Economics and Finance · doi:10.1016/j.najef.2018.06.013 · Cited by 376 · cited 1×
“A comparison of different market price prediction approaches is shown in [21], where classification-based approaches usually provide better results, such as [22]–[24], rather than some regression-based competitors.”
BackgroundSupport vector machine for regression and applications to financial forecasting
Trafalis, T. B., & Ince, H. (2000) · Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium · doi:10.1109/ijcnn.2000.859420 · Cited by 319
Frontiers of finance: Evolution and efficient markets
Farmer, J. D., & Lo, A. W. (1999) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.96.18.9991 · Cited by 304 · cited 1×
“market [11].”
Stock price pattern recognition-a recurrent neural network approach
Kamijo, K., & Tanigawa, T. (1990) · 1990 IJCNN International Joint Conference on Neural Networks · doi:10.1109/ijcnn.1990.137572 · Cited by 252 · cited 1×
“Notwithstanding the above components, the research performed in recent years in this field is growing both in terms of literature production and in tools generation [4], [5]; also, an increasing number of studies involves the use of learning approaches like machine learning classifiers and neural network models.”
BackgroundMeasuring and forecasting S&P 500 index‐futures volatility using high‐frequency data
Martens, M. (2002) · Journal of Futures Markets · doi:10.1002/fut.10016 · Cited by 237
Volatility distribution in the S&P500 stock index
Cizeau, P., Liu, Y., Meyer, M., Peng, C.-K., & Eugene Stanley, H. (1997) · Physica A: Statistical Mechanics and its Applications · doi:10.1016/s0378-4371(97)00417-2 · Cited by 207 · cited 2×
“Indeed, our early tests showed that the simplified VGG led us to significantly improve the results, learning stability and execution times, both when compared to the standard VGG-16 and ResNet [32] architectures.”
ResultFinancial time series prediction using a dendritic neuron model
Zhou, T., Gao, S., Wang, J., Chu, C., Todo, Y., & Tang, Z. (2016) · Knowledge-Based Systems · doi:10.1016/j.knosys.2016.05.031 · Cited by 183 · cited 2×
“Several approaches have also been proposed in the field, facing the financial series prediction topic by using the dendritic neuron model (DNM) technique [9], whose performances have been shown both on Asiatic [10] and U.S. market [11].”
MethodDeep imitation learning for autonomous vehicles based on convolutional neural networks
Kebria, P. M., Khosravi, A., Salaken, S. M., & Nahavandi, S. (2020) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2019.1911825 · Cited by 161 · cited 1×
“As reported in [25], moreover, “there is no general consensus on best forecasting technique for price prediction”, particularly since “price series is inherently a non-stationary series having non-constant mean and variance”, making it not always advantageous to represent the problem through linear models such as regr…”
BackgroundDeep learning for source camera identification on mobile devices
Freire-Obregón, D., Narducci, F., Barra, S., & Castrillón-Santana, M. (2018) · Pattern Recognition Letters · doi:10.1016/j.patrec.2018.01.005 · Cited by 149 · cited 1×
“Indeed, CNNs has shown very good accuracy results when applied to pattern recognition on image and video data [26], [27].”
MethodSelf-citationBrain-inspired Genetic Complementary Learning for Stock Market Prediction
Tan, T. Z., Quek, C. H., & Ng, G. S. (2005) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec.2005.1555027 · Cited by 69 · cited 1×
“This is mainly due to the fact that stock prices are often influenced by factors which are quite hard predictable like political events, the behaviour of the other stock markets and, last but not least, the psychology of the investors [2]; these aspects tend to model the market as an entity which is dynamic, non-linea…”
BackgroundPrice forecasting using wavelet transform and LSE based mixed model in Australian electricity market
Aggarwal, S. K., Saini, L. M., & Kumar, A. (2008) · International Journal of Energy Sector Management · doi:10.1108/17506220810919054 · Cited by 23 · cited 1×
“A comparison of different market price prediction approaches is shown in [21], where classification-based approaches usually provide better results, such as [22]–[24], rather than some regression-based competitors.”
Support Tensor Machine for Financial Forecasting
Calvi, G. G., Lucic, V. M., & Mandic, D. P. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683383 · Cited by 17 · cited 2×
“In a successive step, we applied the RGB color map to the image, thus resulting in a three channel matrix (further details on this process can be found in [28] and [30]).”
BackgroundMethodLee, C. H., & Park, K. C. (1992) · Elsevier eBooks · doi:10.1016/b978-0-444-89488-5.50171-8 · Cited by 14 · cited 1×
“In the same year, the authors in [6] built a recurrent neural network for stock price patterns’ recognition, exploiting the triangle pattern as a clue to the trend of the future stock prices.”
BackgroundSelf-learning of multivariate time series using perceptually important points
Lintonen, T., & Räty, T. (2019) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2019.1911777 · Cited by 14 · cited 1×
“Notwithstanding the above components, the research performed in recent years in this field is growing both in terms of literature production and in tools generation [4], [5]; also, an increasing number of studies involves the use of learning approaches like machine learning classifiers and neural network models.”
BackgroundJia, D., Zheng, S., Yang, L., Todo, Y., & Gao, S. (2018) · IEEE International Conference on Cloud Computing and Intelligence Systems (CCIS) · doi:10.1109/ccis.2018.8691298 · Cited by 13 · cited 1×
“Several approaches have also been proposed in the field, facing the financial series prediction topic by using the dendritic neuron model (DNM) technique [9], whose performances have been shown both on Asiatic [10] and U.”
Method
Cited by
A novel graph convolutional feature based convolutional neural network for stock trend prediction
Chen, W., Jiang, M., Zhang, W.-G., & Chen, Z. (2021) · Information Sciences · doi:10.1016/j.ins.2020.12.068 · Cited by 281
Sonkavde, G., Dharrao, D. S., Bongale, A. M., Deokate, S. T., Doreswamy, D., & Bhat, S. K. (2023) · International Journal of Financial Studies · doi:10.3390/ijfs11030094 · Cited by 262
“We believe that researchers can explore the possible usage of state-of-the-art deep learning and machine learning algorithms in stock trend prediction and classification (Jiang 2021; Nikou et al. 2019).”
BackgroundMulti-DQN: An ensemble of Deep Q-learning agents for stock market forecasting
Carta, S., Ferreira, A., Podda, A. S., Reforgiato Recupero, D., & Sanna, A. (2021) · Expert Systems with Applications · doi:10.1016/j.eswa.2020.113820 · Cited by 226
Self-citationNILM applications: Literature review of learning approaches, recent developments and challenges
Angelis, G.-F., Timplalexis, C., Krinidis, S., Ioannidis, D., & Tzovaras, D. (2022) · Energy and Buildings · doi:10.1016/j.enbuild.2022.111951 · Cited by 219
Integrated deep learning method for workload and resource prediction in cloud systems
Bi, J., Li, S., Yuan, H., & Zhou, M. (2021) · Neurocomputing · doi:10.1016/j.neucom.2020.11.011 · Cited by 162
Metadata
- Authors
- Barra, S., Carta, S. M., Corriga, A., Podda, A. S., & Recupero, D. R.
- Affiliations
- University of Cagliari (IT)
- ORCID
- Venue
- IEEE/CAA Journal of Automatica Sinica
- Published
- 2020-05
- Volume
- 7
- Issue
- 3
- Pages
- 683-692
- ISSN
- 2329-9266 (ISSN-L)
- 2329-9274
- OpenAlex
- W3023384820
- Cited by
- 267
- References
- 30
- Citations by year
- 2026: 27 · 2025: 65 · 2024: 58 · 2023: 36 · 2022: 42 · 2021: 23 · 2020: 16
- FWCI
- 18.61
- Type
- journal article
- Language
- en
- Open access
- green
- hdl.handle.net
- Licence
- other-oa
- Locations
- IEEE/CAA Journal of Automatica Sinica (publishedVersion)
- UNICA IRIS Institutional Research Information System (University of Cagliari) (submittedVersion, other-oa)
- UNICA IRIS Institutional Research Information System (University of Cagliari) (submittedVersion)
- Topics
- Social Sciences › Decision Sciences › Management Science and Operations Research › Stock Market Forecasting Methods
- Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
- Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Complex Systems and Time Series Analysis
- Keywords
- financial forecasting, Gramian Angular Field, ensemble convolutional neural networks, stock trend prediction, S&P 500, multi-resolution imaging, trading systems, buy-and-hold strategy
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.