Research on the Stock Time Series Data Similarity Based on SAX
Weiwei Fu · Computer Engineering and Science · 2009
Research of efficient similarity measurement methods on specific data sets is one of the key research contents in time series data mining.To solve the problem that stock data lack the dynamic information of trend after reducing the dimension by using the SAX method,this paper presents a new similarity measurement function,the Complex-Distance-Function,which joins the point-distance advantages and the model-distance advantages together.Through the experiments of SAX with different distance functions,we prove that the Complex-Distance-Function is useful and provides new ideas to revealing the interdependence between stock data and solve the problem of time series similarity.