Study On EEMD Decomposition Of CPI And Its Fluctuation Period

In a market economy, market regulation plays a fundamental and even decisive role in the optimal allocation of resources. The price mechanism as a core market mechanism, can promote the role of market regulation. As a more sensitive and effective market regulation mechanism, the price mechanism plays an important role in the national economy, and while price stability is one of the four objectives of macroeconomic policy. China currently has achieved the transition of the price mechanism, the industry’s products and services have been gradually liberalized prices. With the pace of economic globalization accelerated, the combined effect of complex economic factors at home and abroad to China’s price volatility has gradually deepened. CPI as the most important measure of price fluctuations, it is not only the characteristics and laws of the cycle has been a focus of academic research, but also has important practical significance.Combing through reading relevant literature and the results of current research and CPI inflation, the paper summarizes the field on the CPI time series characteristics and influencing factors research. Among them, the market supply and demand structure and structural factors multivariate analysis model to CPI fluctuation related modeling, the model itself is a clear economic implications, but because of the price fluctuations of the complexity of factors, in particular multi-scale spatial and temporal characteristics and nonlinear fitting results and the accuracy of the prediction model is not ideal; contrast, data-driven type of short-term forecasting model for the effect of inflation is acceptable, but often lack the analysis of price fluctuations inherent law, the economic implications are not clear. In this regard, the paper chosen Ensemble Empirical Mode Decomposition(EEMD) method, the decomposition of the fixed base index of monthly CPI time series data to obtain a plurality of frequencies independent of sequence. Combined with a new non-parametric identification methods and models of forecasting methods, a new perspective to study China’s CPI fluctuation cycle, and explore the impact of different factors and the internal laws of time and frequency scales CPI fluctuations.Firstly, using EEMD method of decomposing a given base monthly CPI data since 1983, and the components of intrinsic mode function(IMF) and the remaining term trend decomposition obtained were analyzed to clarify the factors on different time scales fluctuations in the CPI impact. Then, using a non-parametric method based on Markov chain, identify the peak date since China’s reform and opening up the monthly inflation rate. Then, combined with the low-frequency part of the decomposition results EEMD intrinsic mode function components corresponding period, the price fluctuations divided since China’s reform and opening up to eight cycles, and the various stages of the background, characteristics, inherent laws and economic factors affecting the analysis. Finally, a method for predicting the CPI for a new attempt, using factors of PROBIT model predictions based on the CPI trend turning point- by calculating the probability of CPI rise or fall, confirmed CPI inflection date, make CPI forecast has practical significance.This article uses qualitative and quantitative analysis of a combination of research periodic variation of the CPI, and the impact of factors at different time scales, while the analysis of a series of macro-control effect since the reform and opening up of the price fluctuations, conclusion has obvious practical significance. However, both the method of choice from the point of view or from the perspective of price fluctuations characterization of view, this type of method(EEMD) still need to be improved. At the same time, we also need to continue to try new methods of analysis in order to better understand the price fluctuations of the relevant characteristics, factors and inherent laws, so as to provide more valuable reference for further research, as well as macro-control policies.