Research on the Application of Convolutional Neural Network in Stock Market Forecasting

Abstract

With the continuous development and complexity of financial markets, stock market forecasting has become a hot topic of concern for investors and decision-makers. This study aims to explore the application of Convolutional Neural Networks (CNN) in stock market prediction, and evaluate their performance in stock index prediction by comparing them with traditional methods such as BP Neural Networks (BPNN) and Support Vector Machines (SVM). In order to further improve the performance of the model, we introduce a model that combines CNN and SVM(CNN-SVM). Experiments show that CNN-SVM model has achieved remarkable advantages in accuracy and robustness, showing the potential value of combining deep learning with traditional machine learning methods. In the empirical analysis, it is found that using CNN-SVM to predict the stock market can achieve relatively superior performance. CNN-SVM can better capture nonlinear relations and complex patterns in time series data through its excellent feature extraction ability. Compared with traditional methods, CNN-SVM shows higher accuracy, precision and recall rate, which verifies its effectiveness in the financial field. The results of this study not only provide investors with more accurate stock market forecasting tools, but also provide a new perspective for the academic research on the combination of deep learning technology in the financial field.