Deep Learning Methods for Stock Price Prediction

Abstract:
The stock market is dynamic and volatile. Prediction of trend of the stock market is considered as important but also difficult. The changes in the stock market are based on numerous factors which are impossible to process such a large amount of data manually and get an accurate result. Therefore, when facing this difficult situation, artificial intelligence is very necessary. There are a lot of methodologies such as Bayesian model, Support Vector Machine classifier, Neural Network, Artificial Neural Networks, Machine Learning Methods and so on, based on the prediction of the stock market. The work of this article is to summarize and classify the past machine learning methods used for stock market prediction, and analyze the advantages, disadvantages and performance of each machine learning method one by one, and under which situation, which machine learning method is used and which method is the most appropriate. At the same time, put forward opinions on the future research trend of stock market changes. From this article, we can conclude that the establishment of a reasonable stock market price prediction model needs to deal with many aspects of complex and huge data. This process is very difficult and challenging. It is very necessary to develop a model that can consider more external complex factors and have more operational efficiency.