Stock trend prediction based on multi-factor screening machine

With the rapid development of financial markets, stock price prediction, as a highly complex and challenging topic, has attracted more and more attention from academia and industry. This study aims to use multiple machine learning algorithms to predict stock market prices, especially exploring the advantages of deep learning models in capturing market nonlinear and time series characteristics. In this article, a machine learning model is used for contribution analysis, analyzing multiple features of stocks, performing a weighted combination analysis of features, and then using a convolutional neural network to perform feature extraction and analysis of stocks, and combining the attention mechanism for deep feature extraction, and finally use the long short-term memory network (LSTM) model to output the model to analyze and predict future stock trends. The experimental results show that combining deep learning analysis technology and factor screening machine can significantly improve the prediction accuracy of the model. At the same time, this article also compared the performance of different models in different market environments, and concluded that in markets with higher volatility, models based on deep learning are more stable. Ultimately, this research provides a new technical means for risk management and investment decisions in financial markets.