DỰ BÁO GIÁ CỔ PHIẾU TRÊN THỊ TRƯỜNG CHỨNG KHOÁN THÔNG QUA MÔ HÌNH PHỨC HỢP LSTM – GRU HYBRID

Abstract

Stock prices are highly complex, nonlinear data, influenced by numerous factors, making the prediction of stock price indices a challenging task. In Vietnam, there are currently very few complex models using machine learning and deep learning, designedwith Python and its available support packages, for forecasting economic variables. This  tudy focuses on a comparative evaluation of deep learning models, including LSTM, GRU, and their hybrid models, to forecast the Vn-Index on the Ho Chi Minh Stock Exchange from January 1, 2009, to January 1, 2024. The evaluation results of the hybrid deep learning model show the lowest deviation from actual values.
Hence, it is recommended that businesses, investors, and policymakers should enhance the use of technology in forecasting to support decision-making processes.