Application of computational verb theory to analysis of stock market data

Abstract

In this paper, computational verb theory (CVT) is applied to the analysis of stock market data. By using CVT, stock market data are clustered into different categories and represented by typical curves for each category. In this paper, researches on the market data samples from Shanghai Stock Exchange in March 2010 are reported. Firstly, MATLAB programs are used to preprocess the stock data. The preprocess consists of curve smoothing, which is achieved by low-pass filtering, and normalization. Secondly, computational verb similarities are used to process the smoothed time series by comparing them with standard computational verbs. Thirdly, Kmeans clustering algorithm is used to cluster the stock data and yields the most representative curves in the stock market.