Research On Earnings Prediction Of Pharmaceutical Industry Based On RBF And LSTM Method

With the development of the national economy,the pharmaceutical industry has been in a good development trend,and the stock prices of pharmaceutical companies in the capital market have also been favored by investors and securities institutions due to their large increase,occupying an important position in the cradle of bull stocks’s position.Therefore,the research on the earnings forecast of the pharmaceutical industry is of great practical significance,whether from the perspective of the company itself,from the perspective of investors or securities analysts.This paper investigates the current situation of the pharmaceutical industry and Hengrui Medicine’s earnings forecasting accuracy,analyzes seven financial indicators of Hengrui Medicine,and applies machine learning algorithms to the scenario of earnings forecasting.First,this paper uses the method of random forest regression to screen the key financial indicators that have an impact on financial earnings through importance analysis.Secondly,select the earnings data of the pharmaceutical industry companies in the first three quarters of the 11 years from 2008 to 2018 when A-shares were listed,and use the RBF neural network and the LSTM neural network to construct an earnings forecast model to predict the earnings in the fourth quarter.Thirdly,forecast the earnings per share of Hengrui Medicine in the fourth quarter from 2008 to 2018,and compare and analyze the accuracy of the earnings forecast model based on the RBF neural network and the LSTM neural network.Finally,from the perspective of the earnings forecast model,combined with the current financial surplus status of the pharmaceutical industry and Hengrui Medicine,it proposes optimization suggestions for the research model,the pharmaceutical industry and Hengrui Medicine.It has been empirically found that the earnings prediction model based on the RBF neural network is better than the model constructed by the LSTM neural network in predicting the earnings of Hengrui Medicine.At the same time,the forecast accuracy of the two models was compared with that of financial analysts,and it was found that the forecast accuracy of the earnings model constructed by the neural network was higher.