Research On Automobile Sales Forecast Based On Baidu Index

After a decade of high growth,the Chinese auto market has entered the normal period of medium and low speed development.With the upgrading and segmentation of consumer demand,the competition of the domestic automotive market will become more intense.At the same time,a more detailed understanding of the market and the more precise prediction of vehicle sales are of great significance for the operation and management of automobile manufacturers and related industries.In the process of the rapid popularization of the Internet,the activities and decisions of consumers are increasingly dependent on the network.The way to obtain information through search engines has replaced most of the traditional channels such as offline consulting.It has gradually become the most important way for hundreds of millions of consumers to obtain information.The largest market share holder in domestic search engines has been the Baidu search engine,and there are trillions of relevant information,personal preference,purchase demand and focus on Baidu.In the context of big data,the extraction and application of Baidu search data is very important for market prediction.By using the Baidu index and taking the Honda automobile brand as the research object,this paper forecasts the car sales,from the selection of key words,screening and synthesis methods,to the test data,the establishment of regression equation and the prediction of sales volume,and makes a detailed analysis.First,using three methods of direct synthesis,stepwise synthesis and principal component analysis,the estimation equation is established,and the least square regression equation is used.The prediction accuracy is analyzed by the mean absolute error,the root mean square error and the mean absolute percentage error of the equation,and the comparison analysis three is made.The advantages and disadvantages of the synthetic methods are obtained.Second,the single factor estimation equation of the historical data and the key word synthesis index is set up respectively,and the advantages and disadvantages of the different equations are compared and analyzed.Through the study of this paper,it is foundthat the key word synthesis index synthesized by principal component analysis is the explanatory variable of the equation,the residual square sum of the equation,the regression standard difference and the fitting degree are all superior,and the final prediction results are most consistent with the actual observation value;the historical data and the keyword synthesis index in the single factor variation are also the most.When alone as independent variable,it is not as high as two variables as independent variables.In this paper,the extension of the difference between the keyword source and the Baidu search index synthesis method enriches the theory of market prediction based on the Baidu index,and based on the empirical analysis of the automobile industry,it shows that the research method of this research method has strong real forecasting ability for automobile sales and can be used for the automobile brand system.The market prediction of manufacturers and auto industry provides effective reference.