Research On The Application Of Improved Exponential Smoothing Model In M Company's Pharmaceutical Sales Forecast

This thesis studies the sales forecast in the context of M company,a medical e-commerce.Due to the wide variety of products,low stability and insufficient supply,there are problems like low prediction accuracy and inaccurate prediction models in forecasting.In order to improve the prediction accuracy,this thesis analyzes the sales data of the target products and finds that,recent data has great value for forecasting,and the promotions and holiday factors cause small fluctuations.In this context,based on the classical exponential smoothing model,an improved exponential smoothing model with the influence of the changes of promotion and holiday factors is established.In order to minimize the error,the optimal value of the smoothing coefficient is solved by MATLAB,and finally the results are verified and analyzed.The results show that,for a certain kind of products,the improved exponential smoothing model predicts the short-term sales trend well,and is better than the classical model.In general,this thesis analyzes the characteristics of the target products and the factors affecting sales fluctuations,provides a better model compared to historical average method,and improves the prediction accuracy.Based on this conclusion,the target company will purchase reasonably,reduce inventory,and save the total cost of during operation.At the same time,this thesis points out the effect of other factors such as advertising and packaging on sales forecast,providing direction for future research and reference for sales of companies.