Comparison Of Methods for Predicting Changes in Stock Trading Volume

Abstract

Stocks occupy an important place in the financial sector, while machine learning, which uses data to solve problems, has been applied in several fields. In this paper, three methods of multivariate linear model, BP neural network and ELMAN neural network are adopted to predict the transaction volume change value of ZTE and compare the evaluation. The study shows that, in the short-term prediction of BP neural network and multiple linear model, the two methods have good prediction effect, the BP neural network is stronger; ELMAN neural network is worse, it is not recommended to predict the trading volume change value.