Predictive Analysis of CPI Based on Machine Learning Algorithms

In recent decades, with the vigorous development of the national economy, the living standards of residents have also risen, and various types of consumption have emerged. In order to study the consumption situation of residents, the government has proposed the Consumer Price Index (CPI) as an indicator to reflect changes in their consumption and service levels. In order to better study the Consumer Price Index, this paper proposes a predictive analysis method for CPI based on machine learning algorithms. The first stage of this research method involves data visualization, which includes two small steps: data preparation explanation and chart visualization. This stage mainly provides initial basis for subsequent modeling; The second stage is the highlight of machine learning (random forest) algorithm modeling, which uses the monthly indices of national, city, and rural levels to train the model separately, with scores of $0.9927,0.9936$, and 0.9919 respectively, and draws the best fitting line; The third stage is for the model to predict the new data for the next month. Through the verification of the entire process mentioned above, the model in this article has good predictive performance and can effectively help solve the problem of CPI prediction.