Research On Mobile Phone Industry Demand Forecast Based On Internet Search Index

In recent years,with the rapid development of mobile phone industry,the supply chain system is becoming more and more mature.The new mobile phone brands and manufacturers resumed their products into the mobile phone market with the support of capital,intensified market competition.During the product development process,an accurate market feedback is required,especially the sales volume forecast in the future.The sales forecast will influence with the supply chain strategy of the mobile phone manufacturer,production capacity adjustment,sales channel planning,etc.It also affects the company's strategic planning.Demand forecasting in the mobile phone industry is an important research direction and has very high practical guiding significance.How to accurately grasp the user's consumption habits and understand the user's real needs is very important in the mobile phone sales forecast.In the Internet era,users' consumption habits have quietly changed,and consumer purchase theories have also continued to develop.At present,the theories have developed to the SICAS model of the mobile Internet stage.The model pointed out that consumers are increasingly relying on social networks and search engines to obtain relevant product information before making purchase decisions.Web search has become one of the important references for measuring whether consumers have the willingness to purchase,especially for consumer electronics product-mobile phone.How to use these web search data for demand forecasting is a new research field,and need to discover the value behind data through data mining technology.The research direction of this paper is the demand forecast of the mobile phone industry based on the web search index.By introducing the typical type of big data for consumer search web keywords' data(using Baidu Index as an example)into the demand forecast of the mobile phone industry.Has been found after introduce the web search data,the mobile phone sales forecasting is more accurate than traditional forecasting models.This paper searches and sorts out the relevant literature of the subject,studies the application of the web search index in various industries,determines to introduce the Baidu Index into the demand forecast of the mobile phone industry,and validates the model's prediction through the example application of Huawei's mobile phone sales in the domestic market.First,the Baidu Index keywords data related to Huawei's mobile phone sales are obtained through web crawlers.The time span is from January 2015 to December 2019.At the same time,the monthly mobile phone sales of Huawei mobile phones is also obtained from the Qianzhan database.According to the correlation analysis and time difference correlation analysis result,9 Baidu Index keywords with guiding significance and a correlation coefficient greater than 0.5 were selected for the study in this paper.Then establish two kinds of sales forecasting models,multiple linear regression and BP neural network,and 1 traditional time series models used to compare.Perform data experiments on the these models,select the first 57 months(a total of 60 months of data set)data to train the model and the next 3 months of data to test the model.The experimental results found that the multi-linear prediction error MAPE value is 6.8%.The prediction error MAPE value of BP neural network is 5.6%,and the prediction error MAPE value of time series is 8.4%.The BP neural network has the best performance.Through research,this paper finds that Huawei's mobile phone sales forecasting model after the introduce the Baidu Index data has improved to varying degrees,proving that the Baidu Index is also suitable for the mobile phone industry demand forecasting,enriching the relevant research on mobile phone demand forecasting.It is of practical guiding significance for mobile phone manufacturers to carry out actual mobile phone sales forecasting applications.