Study On The Methods And Applications For Product Sales Forecasting Based On Online Reviews And Relate Information

With the development of market economy trade and technology,as well as the constant changes in consumer demand,market competition is becoming more and more fierce.Product sales forecasting is an indispensable step in the daily operation of the company.It can help enterprises to consolidate market positioning and provide decision support in inventory management through accurate forecasting of product sales,so that enterprises can better grasp and respond to consumers’ products.At the same time,with the continuous development of the times,the Internet has become an important tool in human daily life.More and more forms of transactions have been switched from offline to online.Most consumers will search for information related to products or post their own comments on the relevant website.In the study of forecasting problems that do not consider the impact of Internet information on product sales,researchers usually use historical sales data of products and attribute indicators of some products to predict.In view of this,it is necessary to further explore the research framework of product sales forecasting methods based on online reviews and related information,and study the corresponding product sales forecasting methods.The new forecasting method not only brings innovation in the research field,opens up new ideas,helps theoretical research integrate with the times and market real economic issues,but also combines theory with practice to provide better forecasts for enterprises in actual market operations.In this paper,the product sales combined forecasting problem based on online reviews and related information is deeply studied.Based on the shortcomings of the existing research,the main research work is carried out from the following aspects:(1)Proposed a single forecasting method for product sales based on online reviews and related information.By analyzing the impact of online commentary and its index data and product network search data on product sales,this paper proposes an online emotional analysis method based on extended sentiment dictionary and BPI index calculation method.According to the characteristics of Bass prediction model,a single forecast model of product sales based on online reviews and related information is constructed.(2)Proposed a combined forecasting method for product sales based on online reviews and related information.Considering the correlation between multiple data and predicted values,and the influence of single model linear features and nonlinear features on product sales forecast results,the impact of combined forecasting methods based on online reviews and related information on product sales is analyzed.According to the Bass-BP prediction model,a product sales combined forecasting model based on multi-source network data is constructed.(3)Given the application research of product sales forecasting method based on online reviews and related information.Taking the product sales forecast of Volkswagen LaVida as an example,the rationality and practicability of the proposed method are verified.The experimental results show that the online reviews and related information can affect the sales of the product,and the accuracy of the combined prediction model is higher than that of the single prediction model.Using a combined forecasting method that combines multi-source network data,it is possible to predict product sales more accurately.