{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,21]],"date-time":"2026-10-21T16:38:35Z","timestamp":1792600715870,"version":"4.5.0"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":52,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>This paper evaluates the forecasting performance of statistical, machine learning, deep learning, and pretrained time-series models across global markets. Using over 10,000 assets spanning equities, bonds, and currencies, we compare ARIMA, XGBoost, LSTM, TimeGPT, and TimesFM under a rolling-origin pseudo-out-of-sample framework. TimesFM records the highest cross-market robustness index, whereas TimeGPT does not outperform the conventional benchmarks on this stability measure. Predictive accuracy is asset-class dependent: LSTM records the lowest RMSE for equities, while XGBoost records the lowest RMSE for bonds and currencies. Overall, the evidence indicates that model performance is conditional rather than uniformly improved by pretraining.<\/jats:p>","DOI":"10.2139\/ssrn.7567225","type":"posted-content","created":{"date-parts":[[2026,10,8]],"date-time":"2026-10-08T04:17:11Z","timestamp":1791433031000},"source":"Crossref","is-referenced-by-count":0,"title":["Do Pretrained Models Improve Financial Time Series Forecasting?"],"prefix":"10.2139","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1426-6593","authenticated-orcid":true,"given":"Xuan","family":"Feng","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"issue":"6","key":"ref1","doi-asserted-by":"crossref","first-page":"1441","DOI":"10.1007\/s00521-013-1386-y","article-title":"A combination of artificial neural network walk models for financial time series forecasting","volume":"24","author":"R Adhikari","year":"2014","journal-title":"Neural Computing and Applications"},{"key":"ref2","first-page":"1","article-title":"Deep learning for stock prediction using numerical and textual information","author":"R Akita","year":"2016","journal-title":"2016 IEEE\/ACIS 15th International Conference on Computer and Information Science"},{"issue":"1","key":"ref3","first-page":"60","article-title":"FinSentGPT: A universal financial sentiment engine","volume":"94","author":"A M Ardekani","year":"2017","journal-title":"International Review of Financial Analysis"},{"key":"ref4","doi-asserted-by":"crossref","first-page":"552","DOI":"10.1016\/j.najef.2018.06.013","article-title":"Predicting the direction of stock market prices using tree-based classifiers","volume":"47","author":"S Basak","year":"2019","journal-title":"The North American Journal of Economics and Finance"},{"key":"ref5","author":"A Borovykh","year":"2017","journal-title":"Conditional time series forecasting with convolutional neural networks"},{"key":"ref6","author":"G E P Box","year":"1976","journal-title":"Time series analysis: Forecasting and control"},{"key":"ref7","author":"G E P Box","year":"2015","journal-title":"Time series analysis: Forecasting and control"},{"issue":"2","key":"ref8","doi-asserted-by":"crossref","first-page":"184","DOI":"10.1007\/s005210170010","article-title":"Financial forecasting using support vector machines","volume":"10","author":"L Cao","year":"2001","journal-title":"Neural Computing & Applications"},{"issue":"4","key":"ref9","doi-asserted-by":"crossref","first-page":"120","DOI":"10.4236\/jcc.2025.134008","article-title":"A Two-Stage Framework for Stock Price Prediction: LLM-Based Forecasting with Risk-Aware PPO Adjustment","volume":"13","author":"Q Chen","year":"2025","journal-title":"Journal of Computer and Communications"},{"key":"ref10","doi-asserted-by":"crossref","first-page":"12","DOI":"10.47852\/bonviewFSI52025703","article-title":"Stock price prediction with LLM-guided market movement signals and transformer model","volume":"1","author":"Q Chen","year":"2025","journal-title":"FinTech and Sustainable Innovation"},{"key":"ref11","first-page":"785","article-title":"XGBoost: A scalable tree boosting system","author":"T Chen","year":"2016","journal-title":"Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining"},{"key":"ref12","doi-asserted-by":"crossref","first-page":"4846","DOI":"10.1109\/BigData62323.2024.10825946","article-title":"Stock price prediction using llm-based sentiment analysis","author":"Q Chen","year":"2024","journal-title":"2024 IEEE International Conference on Big Data (BigData)"},{"key":"ref13","author":"Q Chen","year":"2025","journal-title":"Sentiment-aware stock price prediction with transformer and LLM-generated formulaic alpha"},{"key":"ref14","author":"A Das","year":"2024","journal-title":"A decoder-only foundation model for time"},{"key":"ref15","first-page":"403","article-title":"Artificial neural networks architectures for stock price prediction: Comparisons and applications","volume":"10","author":"L Di Persio","year":"2016","journal-title":"International Journal of Circuits, Systems and Signal Processing"},{"issue":"4","key":"ref16","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1257\/jep.15.4.157","article-title":"GARCH 101: The use of ARCH\/GARCH models in applied econometrics","volume":"15","author":"R F Engle","year":"2001","journal-title":"Journal of Economic Perspectives"},{"issue":"4","key":"ref17","doi-asserted-by":"crossref","first-page":"987","DOI":"10.2307\/1912773","article-title":"Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation","volume":"50","author":"R F Engle","year":"1982","journal-title":"Econometrica"},{"issue":"2","key":"ref18","doi-asserted-by":"crossref","first-page":"654","DOI":"10.1016\/j.ejor.2017.11.054","article-title":"Deep learning with long short-term memory networks for financial market predictions","volume":"270","author":"T Fischer","year":"2018","journal-title":"European Journal of Operational Research"},{"key":"ref19","author":"A Garza","year":"2024","journal-title":"TimeGPT: Foundation models for time series forecasting"},{"key":"ref20","first-page":"6645","article-title":"Speech recognition with deep recurrent neural networks","author":"A Graves","year":"2013","journal-title":"IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)"},{"issue":"10","key":"ref21","doi-asserted-by":"crossref","first-page":"2222","DOI":"10.1109\/TNNLS.2016.2582924","article-title":"LSTM: A search space odyssey","volume":"28","author":"K Greff","year":"2017","journal-title":"IEEE Transactions on Neural Networks and Learning Systems"},{"issue":"8","key":"ref22","doi-asserted-by":"crossref","first-page":"10389","DOI":"10.1016\/j.eswa.2011.02.068","article-title":"Using artificial neural network models in stock market index prediction","volume":"38","author":"E Guresen","year":"2011","journal-title":"Expert Systems with Applications"},{"key":"ref23","article-title":"FinLSPM: Large Stock Predict Model via Numerical Prior Knowledge from LLM","author":"H Guo","year":"2025","journal-title":"Expert Systems with Applications"},{"issue":"2","key":"ref24","first-page":"453","article-title":"The model confidence set","volume":"78","author":"P R Hansen","year":"2010","journal-title":"Econometrica"},{"issue":"8","key":"ref25","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"S Hochreiter","year":"1997","journal-title":"Neural Computation"},{"key":"ref26","author":"S A Hussain","year":"2026","journal-title":"Improving Financial Forecasting with a Synergistic LLM-Transformer Architecture: A Hybrid Approach to Stock Price Prediction"},{"key":"ref27","author":"R J Hyndman","year":"2018","journal-title":"Forecasting: Principles and practice"},{"key":"ref28","first-page":"1371","article-title":"Short and long term stock trend prediction using decision tree","author":"R A Kamble","year":"2017","journal-title":"2017 International Conference on Intelligent Computing and Control Systems"},{"key":"ref29","first-page":"3146","article-title":"LightGBM: A highly efficient gradient boosting decision tree","volume":"30","author":"G Ke","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"issue":"2","key":"ref30","doi-asserted-by":"crossref","first-page":"569","DOI":"10.1016\/j.asoc.2006.03.004","article-title":"A hybrid approach based on neural networks and genetic algorithms for detecting temporal patterns in stock markets","volume":"7","author":"H Kim","year":"2007","journal-title":"Applied Soft Computing"},{"key":"ref31","article-title":"Time series forecasting with deep learning: A survey","volume":"379","author":"B Lim","year":"2021","journal-title":"Philosophical Transactions of the Royal Society A"},{"issue":"4","key":"ref32","first-page":"1748","article-title":"Temporal fusion transformers for interpretable multi-horizon time series forecasting","volume":"37","author":"Y Li","year":"2020","journal-title":"International Journal of Forecasting"},{"issue":"2","key":"ref33","doi-asserted-by":"crossref","first-page":"383","DOI":"10.1111\/j.1540-6261.1970.tb00518.x","article-title":"Efficient capital markets: A review of theory and empirical work","volume":"25","author":"B G Malkiel","year":"1970","journal-title":"The Journal of Finance"},{"issue":"3","key":"ref34","first-page":"804","article-title":"Statistical, machine learning and deep learning forecasting methods: Comparisons and ways forward","volume":"73","author":"S Makridakis","year":"2022","journal-title":"Journal of the Operational Research Society"},{"key":"ref35","author":"O Mutian","year":"2025","journal-title":"LLM-guided semantic feature selection for interpretable financial market forecasting in low-resource financial markets"},{"key":"ref36","doi-asserted-by":"crossref","first-page":"1419","DOI":"10.1109\/IJCNN.2017.7966019","article-title":"Stock market's price movement prediction with LSTM neural networks","author":"D M Nelson","year":"2017","journal-title":"2017 International Joint Conference on Neural Networks"},{"key":"ref37","doi-asserted-by":"crossref","first-page":"181","DOI":"10.1016\/j.eswa.2019.01.083","article-title":"Combining principal component analysis, discrete wavelet transform and XGBoost to trade in the financial markets","volume":"125","author":"J Nobre","year":"2019","journal-title":"Expert Systems with Applications"},{"key":"ref38","article-title":"N-BEATS: Neural basis expansion analysis for interpretable time series forecasting","author":"B N Oreshkin","year":"2020","journal-title":"International Conference on Learning Representations (ICLR)"},{"key":"ref39","first-page":"3687","article-title":"Data-driven medium-range weather prediction with a resnet and transformers","volume":"14","author":"S Rasp","year":"2021","journal-title":"Geoscientific Model Development"},{"issue":"6","key":"ref40","doi-asserted-by":"crossref","first-page":"3234","DOI":"10.1016\/j.eswa.2014.12.003","article-title":"Recurrent neural network and a hybrid model for prediction of stock returns","volume":"42","author":"A M Rather","year":"2015","journal-title":"Expert Systems with Applications"},{"key":"ref41","author":"M Roondiwala","journal-title":"Predicting stock prices using LSTM"},{"issue":"4","key":"ref42","doi-asserted-by":"crossref","first-page":"536","DOI":"10.1016\/j.jksuci.2015.06.002","article-title":"Forecasting financial time series using a low complexity recurrent neural network and evolutionary learning approach","volume":"29","author":"A K Rout","year":"2015","journal-title":"Journal of King Saud University -Computer and Information Sciences"},{"key":"ref43","doi-asserted-by":"crossref","first-page":"1643","DOI":"10.1109\/ICACCI.2017.8126078","article-title":"Stock price prediction using LSTM, RNN and CNN-sliding window model","author":"S Selvin","year":"2017","journal-title":"2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI)"},{"issue":"1","key":"ref44","doi-asserted-by":"crossref","first-page":"83","DOI":"10.1257\/089533003321164967","article-title":"From efficient markets theory to behavioral finance","volume":"17","author":"R J Shiller","year":"2003","journal-title":"Journal of Economic Perspectives"},{"key":"ref45","first-page":"562","article-title":"Stock Price Prediction with Attention-Based Framework by Integrating LLM-Generated Features","author":"Y Sun","year":"2025","journal-title":"International Conference on Neural Information Processing"},{"issue":"1","key":"ref46","first-page":"19","article-title":"Performance analysis of stock price prediction using artificial neural network","volume":"12","author":"K K Sureshkumar","year":"2012","journal-title":"GJCST"},{"key":"ref47","first-page":"3104","article-title":"Sequence to sequence learning with neural networks","volume":"27","author":"I Sutskever","year":"2014","journal-title":"Advances in Neural Information Processing Systems (NeurIPS)"},{"key":"ref48","doi-asserted-by":"crossref","DOI":"10.1002\/0471746193","author":"R S Tsay","year":"2005","journal-title":"Analysis of financial time series"},{"key":"ref49","first-page":"5998","article-title":"Attention is all you need","volume":"30","author":"A Vaswani","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref50","first-page":"759","article-title":"Predicting stock market trends by recurrent deep neural networks","author":"A Yoshihara","year":"2014","journal-title":"Proc. Pacific Rim International Conference on Artificial Intelligence"},{"key":"ref51","first-page":"1233","article-title":"A review of recurrent neural networks","volume":"31","author":"Y Yu","year":"2019","journal-title":"Neural Computing and Applications"},{"issue":"3","key":"ref52","doi-asserted-by":"crossref","DOI":"10.3390\/math13030487","article-title":"Llm-augmented linear transformer-cnn for enhanced stock price prediction","volume":"13","author":"L Zhou","year":"2025","journal-title":"Mathematics"}],"container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,21]],"date-time":"2026-10-21T15:40:59Z","timestamp":1792597259000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7567225"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":52,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7567225","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}