Convex Optimization for Explainable Machine Learning: A Sparse Regularization Perspective
DOI:
https://doi.org/10.54097/6k9p2014Keywords:
Convex optimization, explainable machine learning, Stock market forecast.Abstract
In recent years, the use of machine learning techniques for predicting the stock market has grown quickly. However, more and more people concern model explainability since the "black box" model problem has come up as a result. This essay examines the impact of convex optimization on explainable machine learning from the standpoint of mathematical optimization. It shows that Lasso ( ) and Ridge ( ) regularized objectives are convex, and that adding makes them strongly convex. Thus, the unique global optimal solution is obtained. This research use Apple's stock data from 2019 to 2024 to build sparse regression models like the Lasso, Ridge, and Elastic Net. Then this research compares how well those models predict and how well the model explains the feature. Finally, results show that Elastic Net strikes the best balance between explainability and generalization. It better than Lasso and Ridge in walk-forward MSE. Also, it keeps coefficients that are stable and economically meaningful. In general, using convex optimization and sparse regularization makes models so that it can be proven mathematically and explained in terms of economics. This is a principled way to make AI in finance more transparent and reliable.
Downloads
References
[1] Molnar, C. Interpretable machine learning: A guide for making black box models explainable (2nd ed.). Lulu Press, 2022.
[2] Rudin C. Stop explaining black box machine learning models for high-stakes decisions and use explainable models instead. Nature Machine Intelligence, 2019, 1(5), 206–215. https://doi.org/ 10.1038/s42256-019-0048-x
[3] Tiddi I, Schlobach S. Explainable artificial intelligence (XAI): A systematic review of methods and applications in finance. Artificial Intelligence Review, 2023, 57(2), 1679–1711. https://doi.org/ 10.1007/s10462-024-11077-7
[4] Roscher R, Bohlender J, Kieseberg P. Balancing accuracy and interpretability: Empirical evaluation of generalized additive models for explainable AI in finance. Business & Information Systems Engineering, 2025, 67(1), 45–58. https://doi.org/10.1007/s12599-024-00922-2
[5] Gu S, Kelly B, Xiu D. Empirical asset pricing via machine learning. The Review of Financial Studies, 2020, 33(5), 2223–2273. https://doi.org/10.1093/rfs/hhaa009
[6] Bergmeir C, Hyndman R J, Koo B. A note on the validity of cross-validation for evaluating time series forecasting models. Journal of Time Series Analysis, 2018, 39 (6), 1262–1264. https://doi.org/ 10.1111/jtsa.12412
[7] Cerqueira V, Torgo L, Mozetič I. Evaluating time series forecasting models: An empirical study on performance estimation methods. Machine Learning, 2020, 109(11), 1997–2028. https://doi.org/ 10.1007/s10994-020-05950-0
[8] Tibshirani R. Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B, 1996, 58(1), 267–288.
[9] Zou H, Hastie T. Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society: Series B, 2005, 67 (2), 301–320. https://doi.org/10.1111/j.1467-9868.2005.00503.x
[10] Hastie T, Tibshirani R, Wainwright M. Statistical learning with sparsity: The lasso and generalizations. CRC Press, 2015.
[11] Hyndman R J, Athanasopoulo G. Forecasting: Principles and practice (3rd ed.). OTexts, 2021.
[12] Xu Y, Chen X, Wang L. Convex optimization methods for explainable sparse modeling. IEEE Transactions on Neural Networks and Learning Systems, 2021, 32(11), 4975–4990. https://doi.org /10.1109 /TNNLS.2020.3043497
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
How to Cite
