The Application Of Frequency Domain Analysis In Chinese Stock Market

The business cycle is an important economic theory,this article analyzed the cycle of Chinese stock market with frequency domain analysis,then build a cyclic model with the result.For the correlation of Chinese stock market and international market,this article applied Fourier Transform and MUSIC algorithm on S&P500,CSI 300 and Nikkei 225,find that,for American and Japanese market,there exists four main frequency components,400 months,120 months,46months and 28 months.For Chinese market,there exists three main frequency components,170 months,48 months and 30 months.For the correlation of Chinese stock market and economy,this article applied Fourier Transform and MUSIC algorithm on CPI and M2,find out that,they have two main frequency components similar to stock market.After that,we applied wavelet analysis to Chinese stock market,find out that,the three main frequency components keep consistency on time.Based on the result found,this article constructed a cyclic model with three main frequency components.The model can explain over 60% change of stock market.After that,this article build a trading strategy with its ability of prediction,the strategy has an annualized return of 14.1%.