The Influence of Volume and Volatility on Predicting Shanghai Stock Exchange Trends

摘要

Most of the previous studies concerning mining association rules from stock time series simply use confidence and support thresholds. In this paper we introduce two new thresholds -– trading volume and stock volatility -- that suit stock time series behaviour better. In this study, we test the influence of volatility and volume on share price weekly trends. Various experimental results yield the strong correlation between trading volume and classifying accuracy. We use the mined rules to classify and predict future trends. A new method, namely weighted confidence, is proposed for carrying out associative classification/ prediction. Its accuracy is equivalent to other traditional measures.