ESTIMATION AND PREDICTION OF CORE INFLATION

Through comparing 30% one-sided trimming approach, structural vector autoregression (SVAR) approach, variance-weighted index approach, HP filter approach, persistence-weighted approach and variance trimming approach (variance symmetrical trimming and variance one-sided trimming), it is found that variance trimming approach is more suitable to predict the trend of inflation in China. Hence, this method is used to estimate core consumer price index (core CPI) from January 2001 to September 2013. In addition, nonparametric support vector regression is used to predict the inflation trend of China from October 2013 to March 2014. It is concluded that China is still in the stage of inflation in next six months, but the level of inflation will decline.