Comparative Analysis of Prophet and LSTM Model in Drug Sales Forecasting

Abstract

With the rapid development of China's pharmaceutical industry, accurate prediction of drug sales has become the key to enterprises' competitiveness. Sales forecasting research has very important value for strategic decisions and improvement measures made by enterprises. We studied mainly two machine learning methods of pharmaceutical sales prediction in this paper, analyzed deeply the prophet model and LSTM, and carried out a comparative experiment on these two methods using real sales data. The experimental results show that the LSTM model is more accurate than the Prophet model on drug sales forecasting.