Sales Forecasting Model of E-commerce Activities Based on Improved Random Forest Algorithm

The accuracy of commodity sales forecast is related to the profits of all stakeholders. But shortages and overstocking are widespread, posing a dilemma for retailers: They need to balance losses from shortages with the cost of safe stocking. This paper mainly studies the data mining algorithm for mobile e-commerce activity sales forecasting model. This paper first introduces the concept of data mining, and improves the random forest algorithm in the data mining algorithm, and puts forward the method of Bayesian optimization combined with time series segmentation to optimize the random forest model, so as to achieve the purpose of improving the prediction effect. Based on the improved random forest algorithm, this paper established the sales prediction model of mobile e-commerce. Through the analysis of the results of the prediction model, it can be seen that the prediction error of the prediction model is larger for food and clothing whose sales are greatly affected by activities, and smaller for tools whose sales are less affected by activities.