E-commerce Retail Sales Trend Analysis and Prediction Based on the ARIMA Model

E-commerce retail sales have grown significantly recently due to technological advancements and the COVID-19 pandemic. The importance of accurate sales forecasting in this domain is boosted drastically. The study addresses a gap in the field regarding the robustness of forecasting models in the face of abrupt market shifts. This paper delves into the prediction of e-commerce retail sales trends using the Autoregressive Integrated Moving Average (ARIMA) model and analyzes the COVID-19 pandemic’s unprecedented impact on consumer behavior and the digital economy. The research aims to evaluate the ARIMA model’s efficacy in forecasting e-commerce sales in the current market. This research employs two data-splitting methods to assess the impact of the pandemic on prediction accuracy. The study reveals that while the ARIMA model performs well in stable periods, it struggles with the volatility and unpredictability introduced by the pandemic. The model’s predictions for the post-pandemic period show significant deviations from actual values. The research concludes that despite the ARIMA model’s utility in short-term forecasting, it requires enhancement during crises like the COVID-19 pandemic.