Monte carlo tree search improves sparse feature selection for financial regression
$L_0$ Feature Selection via Monte Carlo Tree Search in High-Dimensional Financial Regression
Machine Learning
Summary
Picking the right key variables is important when making sense of lots of financial data, but usual methods struggle with variables that are closely related. The authors use a method called Monte Carlo Tree Search to try out different groups of features, focusing on how well they explain new data instead of just fitting old data. This approach found smaller sets of variables that better explained stock index movements than a popular technique called Lasso. Their approach offers a new way to pick important features with better clarity and accuracy in financial studies.
What this means in practice
- •For financial analysts: Identify compact sets of explanatory variables to better interpret financial index returns with improved out-of-sample clarity.
- •For risk management teams: Decompose risk factors more transparently by selecting feature subsets that account for variable correlations without shrinkage bias.
Tested on one dataset.
Authors
Cameron Mouangue, Tristan Cazenave
Abstract
Feature selection is a central problem in high-dimensional financial regression, where large numbers of engineered variables and strong correlations among predictors reduce model stability and interpretability. $L_1$ regularization (Lasso) is widely used for sparse regression, but its shrinkage can be restrictive when informative signals are spread across correlated variables. Exact sparse regression avoids this shrinkage but requires solving a combinatorial optimization problem that is computationally difficult and is typically fit to minimize in-sample rather than out-of-sample error. This paper presents a feature selection framework based on Monte Carlo Tree Search with Generalized Nested Rollout Policy Adaptation (GNRPA). Feature selection is formulated as a sequential decision process, and candidate subsets are scored by their out-of-sample validation performance rather than training error. The method also supports a lexicographic objective accounting for subset size and multicollinearity. We study this approach on a contemporaneous explanatory regression task: explaining the weekly return of the NASDAQ-100 index from 1,072 engineered features spanning market, macroeconomic, and factor data. This isolates the value of subset selection under multicollinearity from the separate problem of predictive forecasting. On this task, the proposed method identifies sparse feature subsets with higher out-of-sample explanatory power than an optimized Lasso baseline, reaching an out-of-sample $R^2$ of 0.6961 versus 0.6313 for Lasso. These results suggest that Monte Carlo Tree Search is a practical alternative to convex regularization for sparse explanatory feature selection in financial regression, with potential value for factor attribution and risk decomposition where transparency matters as much as out-of-sample fit.