Advanced Machine Learning for Financial Markets: A PCA-GRU-LSTM Approach

Abstract

This study pioneers the integration of environmental data with financial indicators to forecast stock prices, employing a novel PCA-GRU-LSTM model. By analyzing the Shanghai Composite (SSEC) index alongside six key air pollutants, we illuminate the significant role of environmental factors in financial forecasting. The PCA-GRU-LSTM model, which combines principal component analysis (PCA), gated recurrent units (GRU), and long short-term memory (LSTM) networks, demonstrates superior predictive accuracy by leveraging both financial and environmental datasets. Our findings indicate that incorporating environmental indicators enriches the model’s data set and significantly enhances forecasting precision, especially when adjusted for seasonal variations. This study’s results underscore the potential for more sustainable investment strategies, emphasizing the interconnectedness of environmental and financial systems. By offering insights into the dynamic interactions between environmental variables and stock market fluctuations, this research contributes to the burgeoning field of sustainable finance, urging the inclusion of environmental considerations in financial decision-making processes. The PCA-GRU-LSTM model’s success highlights the importance of leveraging advanced machine learning techniques to capture the complex, multifaceted nature of stock price movements, offering a promising avenue for future research in the knowledge economy’s intersection of technology, innovation, and society.