Time Series Forecasting Model for E-commerce Store Sales Using FB-Prophet

Authors:

Abstract
Recently, there has been a huge increase in online sales. Therefore, increasing e-commerce sales improves the quality of business sales. To increase the efficiency of this process, decision makers in the sales domain need technology that helps them make decisions and reduces the rate of errors. In this paper, we conduct experiments in two main steps: First, we handled missing data in the data set using various techniques. Second, we compare two forecasting models, namely ARIMA and FB-Prophet. The results indicate that the ARIMA model outperforms the FB-prophet model in forecasting sales data; besides, the results prove the superiority of the method of handling the missing data with the Forward and Backward functions over processing the data by filling in with zeros.

This publication has 13 references indexed in Scilit: