Analysis and forecast of stock price based on LSTM algorithm

In the 21st century, the stock is a high-risk, high-yield common financial products, in order to better analyze the stock investment, effectively and accurately predict the stock price trend has become the current research hotspot in recent years. In this paper, we put forward a LSTM prediction model to predict the closing stock prices, taking Microsoft as an example. Experimental results indicate that this method displays better prediction performance than other machine learning methods. Specifically, the RMSE value is 24.43% and 16.69% lower than ridge algorithm and neural network algorithm respectively. In addition, we make a detailed data analysis of the data set, including the stock price changes over time, the average daily return and the moving average of the stock, and give how much risk to invest in a specific stock, which can provide some guidance for our future work.