Trend Prediction Analysis of Shanghai Composite Index Based on LSTM Neural Network

Abstract

The prediction of stock indices, particularly the Shanghai Composite Index (SCI), is crucial for understanding the health and direction of China's financial market in the post-pandemic era. Given the challenges of traditional forecasting methods due to the unpredictable nature of stock prices, this study explores the application of the Long Short-Term Memory (LSTM) neural network, complemented by Monte Carlo simulations and regularization with the daily trading data from September 21 2020, to August 13, 2023, to forecast the SCI's price trends for Q4 2023. The findings suggest a trajectory characterized by an initial decline, succeeded by a steady upward trend throughout the entirety of the quarter. Notably, the results have adeptly encapsulated both the lingering effects of the pandemic and the long-term rising trend of the Shanghai Composite Index (SCI). This research accentuates the potential of advanced neural network models in deciphering complex stock market behaviors, offering a groundbreaking perspective for market participants in navigating future investment decisions.