CPI Prediction Based On Time Series Analysis

CPI is an important indicator reflecting the degree of inflation and affecting macroeconomic policies.The prediction of CPI has always been the focus of economics and has rich practical significance.This thesis first introduces the definition of different kinds of CPI data,analyzes their respective advantages and disadvantages,and selects the data of CPI year-on-year and CPI month-on-month for empirical research.After a brief introduction to the basic theory and model of time series,this thesis analyzes the seasonal and trend terms for CPI year-on-year and CPI month-on-month data,then uses the ARIMA model to forecast the CPI data and has obtained good fitting and prediction results.Finally,the performance of each model and different CPI data is analyzed and evaluated,and their advantages and disadvantages and adaptation scenarios are summarized.