Time Series Analysis Based On The Strength Of Stock Capital Flows

The relationship between quantity and price has become one of hot topics in the study of the stock market. The research of relationship between quantity and price include the correlation between strike price and volume, yield and volume, yield and turn volume. Due to the yield is stationary series, most of the research is based on the correlation between yield and volume. But the volume can’t reflect the stock information comprehensively. Strength of stock fund flows(SOSFF) is a new technical indicators to compile for strike price and volume, which can reflect the relevant information of stock market comprehensively. Research of the relationship between SOSFF and yield, and make empirical analysis can reflect the stock trend more objective and accurate, and also achieve the purpose of predicting stock price.The paper regard SOSFF which with five and fifteen as circle as the research object, and makes a comprehensive empirical analysis, and makes descriptive statistics, autocorrelation analysis, stationary test for its time series, establishes the ARMA model. To study the relationship between SOSFF and the yield, the paper make regression analysis and Granger causality test for them. And research if SOSFF can eliminate the leverage effect in the stock market based on the GARCH model. Simultaneously, make a comprehensive empirical analysis for 14 stocks in banking industry of shanghai stock market.The empirical results show that SOSFF is a stationary series with the characteristics of rush thick tail, and has highly serial correlation. There is a significant positive correlation between SOSFF and yield absolute value. Granger causality test results show that there is two-way granger causality between SOSFF and absolute yields. It suggests that SOSFF indeed contains price changes information, and can provide certain help for price prediction. Also find that logarithmic SOSFF can decrease the yield of the ARCH effect. It indicates that SOSFF contains some information about price changes. Relevant empirical results suggest that time series analysis based on SOSFF is an effective method to predict the stock market price.