A Forecast Study Of CPI Based On Web Search Data

The Consumer Price Index(CPI)is an important macroeconomic index that reflects the changes in the price level of various consumer goods and other public service logistics related to the daily life of Chinese residents.It is also an analysis and decision-making of the macroeconomic and the national economy.The important measurement indicators for accounting provide a decision-making basis for the society and the country to formulate macro-control policies.It provides a decision-making basis for the country to formulate macro-control policies;however,there is always a lag period of about half a month when the statistics department releases the CPI.Therefore,the prediction of CPI has also become a research hotspot.The development of global network technology can be said to be changing with each passing day.Coupled with the rapid improvement of the domestic Internet,the way people obtain information is developing from traditional channels to network search engines.What follows is a massive amount of network data,and such massive amounts of data are constantly being generated.If the big data contained in the Internet is transformed into useful information,it will bring huge social and economic benefits.At present,consumers usually use search engines to inquire before purchasing consumer products.The search information of consumer products is recorded in the search engine in the form of web search data.The search information of consumer products can reflect the behavior of consumers in a timely manner,and thus can reflect On the changes in commodity prices.This article first makes a theoretical analysis of the influencing factors of commodity prices and the transmission time lag.Among them,the two most important factors affecting prices are macroeconomic policies and the relationship between supply and demand,and their impact on prices has a certain lag.When the macroeconomic policy and the relationship between supply and demand change,the search volume of specific keywords will correspondingly change in real time.Furthermore,it analyzes the correlation mechanism between the Internet search data and the consumer price index,and discusses the theoretical relationship with keywords from the perspective of macro factors and the relationship between supply and demand.It is believed that there is a certain correlation and lag between the search data and the price level.Selecting relevant keywords and establishing a model can make predictions and judgments on the trend of the price level.Then select the corresponding keywords,crawl the data under the search engines of Baidu Index and Ali Index respectively,and reduce the dimensionality of the data by combining correlation determination and principal component analysis to increase the degree of fit with CPI,And finally synthesize two types of comprehensive indexes into macro and micro.Using these two types of comprehensive indexes as the model data,we will continue to explore the correlation between the comprehensive index and CPI,and we will find that there are lagging factors in the data through analysis.Therefore,the regression model,the lag model and the neural network model are established respectively.Through the comparison of the models,it is concluded that considering the influence of the lag factor can increase the fitting ability of the final model,and the root mean square error and the average absolute error of the lag model and the neural network model are both Smaller.The final result shows that the CPI fitted by the neural network model is consistent with the real CPI trend.The root mean square error and average absolute error of the model are the smallest,and there is only a 0.034 difference between the predicted value and the real value.Therefore,the method of predicting CPI in this article can overcome the lag of traditional CPI statistical methods,and can predict the trend of CPI to a certain extent in advance,which can provide a reference for future macroeconomic forecasts.