LSTM-Based Predictive Stock Selection for ESG Portfolio Optimization in the Malaysian Stock Market

View at publisher Summarise with VIRAT
Authors: Rattapon Sriwichupong, Siriratch Sirisomphobchai, Nichapa Sirisakdanusorn, Jeeraporn Ananpraphakorn, Rujira Chaysiri
Cited by: 0
References: 24

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study investigates the financial viability of sustainable investments using a machine learning-based framework to evaluate ESG-screened and conventional portfolios in the Malaysian stock market. A Long Short-Term Memory (LSTM) network is employed for daily return forecasting, which is integrated into three quarterly rebalanced portfolio optimizations: Mean-Variance (MV), Black-Litterman (BL), and Equal Weight (1/N). Results from a one-year out-ofsample test indicate that portfolio performance weakens after accounting for transaction costs. Nevertheless, the ESG portfolio demonstrates greater resilience than the Non-ESG and Mixed portfolios across all approaches. In particular, the ESG portfolio exhibits lower volatility and smaller maximum drawdowns. While the Non-ESG portfolio occasionally achieves higher risk-adjusted metrics, it is associated with larger drawdowns, indicating higher exposure to adverse market conditions. Overall, the results support the role of ESG integration in enhancing portfolio resilience in an emerging market context.

Fields of Study & Subjects

Stock Market Forecasting Methods Sustainable Finance and Green Bonds Financial Markets and Investment Strategies

Publication details

Journal / source
International Conference on Knowledge and Smart Technology
Pages
261–266
Publication date
21 January 2026
Language
EN