Synapse
⌘+K
Synapse
PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
March 25, 2026ElectronicsOpen Access

Regime-Aware LightGBM for Stock Market Forecasting: A Validated Walk-Forward Framework with Statistical Rigor and Explainable AI Analysis

View Full Paper
Ask AI
Bookmark
Share

Authors

APA. Pagliaro

Discussion

Loading...

Member takes

Overview

Demonstrates a regime-aware machine learning model improves stock market predictions, suggesting enhanced adaptability.

Key Points

  • This research aims to evaluate whether a machine learning framework can produce validated returns in varying stock market conditions.
  • Developed a regime-aware LightGBM framework utilizing a Hidden Markov Model for market regime detection.
  • Conducted backtesting on 51 NASDAQ-100 stocks from 2015 to 2026.
  • Eliminated look-ahead bias by implementing a walk-forward approach.
  • Achieved a portfolio Sharpe ratio of 1.18 with a 95% confidence interval of [0.53, 1.84].
  • Identified that cross-asset features like Bitcoin provide significant predictive value.
  • Confirmed macroeconomic indicators are more impactful than technical indicators in high-beta market conditions.

Cite This Study

A. Pagliaro (2026) studied this question.

synapsesocial.com/papers/69c37b54b34aaaeb1a67d9bfhttps://doi.org/10.3390/electronics15061334
View Full Paper
Ask AI
Bookmark
Share