Cross-Sectoral Crash Contagion Forecasting: An Ensemble Learning Framework for Multi-Stock Market Crisis Prediction in Emerging Economies

Predicting coordinated stock market crashes re-mains a challenging problem in computational finance, particularly in emerging economies where market behavior is highly volatile and interconnected. This paper proposes an ensemble learning framework for forecasting multi-stock market crashes using selected Nifty 50 companies from different industrial sectors. The proposed approach combines temporal financial indicators and crosssector relationships using a Random Forest-based predictive model. Historical stock data from 2018 to 2023 were analyzed using time-series cross-validation to avoid data leakage and improve evaluation reliability. Experimental results achieved a prediction accuracy of 67.7% and demonstrated the ability of the model to identify sector-wise crash propagation patterns. The analysis indicates that technology and financial sectors frequently act as leading indicators of broader market stress. The proposed framework provides useful insights for portfolio risk management, systemic risk monitoring, and future financial forecasting research in emerging markets.