The Research of Morphological Characteristics in Time Series of Stock Prices Based on CBR

The paper analyzes technology specific morphological characteristics of securities time series, proposes data structure, and takes the turning point of the short-term trend as the observation point to identify specific technical form to obtain its case data set. By Q cluster analysis based on stocks of nine financial indicators, we build a case library to retrieve similar cases after determining the index weights with the method of multiple linear regression, which can be used to predict the securities trends. The results show the feasibility and effectiveness of the method through empirical analysis of historical data of the Shanghai and Shenzhen stocks.