Stock Daily Return Prediction of Amazon and Alibaba using Linear Regression and LSTM

Abstract

The stock plays a key role in the economy market. As an individual investor, predicting stock return is always a hot research topic. In recent years, Amazon and Alibaba have become the head of e-commerce and the competition between them is becoming intense. In this paper, linear regression model and LSTM model based on machine learning are introduced and applied to predict the stock return of these two companies. RMSE and R2 are used to choose the number of variables and evaluate those models. This study finds that the simplest linear regression is even better than LSTM model with limited sources. The negative R2 scores of these two methods implies the nonlinearity and instability of stock return. However, predicting stock return is still possible. To investigate more about predicting stock return, more information such as the turnover rate and other non-numerical variables can be included in other models. Different types of LSTM model with different parameter setting can be applied to investigate deeply.