Prediction And Empirical Study Of Consumer Confidence Index Based On Internet Search Data

Consumer confidence index(CCI),as a leading indicator to forecast economic trends and consumption trends,measures consumers' subjective feelings of a country's current macroeconomic situation and future prospects.It is an important economic development boom index of a country.The consumer confidence index compiled by China's National Bureau of Statistics is based on sampling survey data,which are obtained by random survey of consumers in more than 20 cities in China in the form of questionnaire.This traditional method of sampling survey is time-consuming,laborious,costly and time-limited.In the era of Internet and big data,social economic behavior and information technology increasingly converge.Massive internet data contains various behaviors and intentions of consumers,which can reflect the social and economic changes and consumers' psychological expectations in real time.Under this background,the thesis constructs a correlation analysis mechanism between Internet search data and consumer confidence index.Moreover,the thesis focuses on the composition and connotation of consumer confidence index,and obtains the related internet search data from Baidu Index platform.In the process of data cleaning,dimension reduction,modeling,verification and analysis,big data mining method and traditional statistical analysis method are organically combined.The empirical research forecasted the current consumer confidence index in real time.It not only improves the forecasting efficiency of consumer confidence index,but also saves a lot of manpower and material cost.