Correlation Analysis of Industry Sectors in China's Stock Markets Based on Interval Data

Wen Long, Yeran Tang, Dingmu Cao

Comparing with single value data, interval data is an important data type containing more information. This paper focuses on correlation analysis of interval data and proposes a comprehensive weighted correlation coefficient which combines both trend and range information of a sequence. Applying the new approach of interval data correlation coefficient to China's stock market, we obtain the empirical value of the weight taking CSI 300 as the representative. Further, the correlations between industry sectors as well as the sample stocks within a specific industry sector are analyzed respectively using this method. The empirical studies suggest that the correlation coefficient of interval data has improved the results of traditional correlation coefficient and the temporal fluctuation characteristics of the sequence are better reflected based on interval data.