Predicting Stock Prices Using Hybrid LSTM and ARIMA Model

This technical report presents a hybrid model that combines Long Short-Term Memory (LSTM) and Auto Regressive Integrated Moving Average (ARIMA) approaches for predicting stock prices. The report outlines the increasing interest of investors in China's financial market and the demand for accurate financial information services. It details the construction of a stock dataset representative of the market and describes the methodology for developing the hybrid model. The report explains the training process of the LSTM model, including the design of the network structure and the selection of activation functions to enhance prediction accuracy. It also discusses the optimization of the model to prevent overfitting and the use of Mean Squared Error as a loss function. The results indicate that the proposed hybrid model demonstrates higher prediction accuracy and stability compared to traditional models, making it a valuable tool for stock trend prediction.