M.-C. Chan and H.C. Leung (PRC)
data mining, association rules, financialratios, China, Hong Kong, stock
Financial statement analysis of listed companies is essential in investment and portfolio management for maximizing investment return with controlled risk. Usually the analysis involves large amount of data and the underlying patterns may not be easy to identify. With the increasing power of computer, the knowledge discovery in database (KDD), i.e., data mining, becomes possible in the domain of finance. The focus of this study is to extract association rules for financial ratios against stock price change. A data mining application for association rules on China and Hong Kong Stock Market has been built and to be accessed through the WEB. Preprocessing has been done to convert the data on financial report to financial ratios to allow for direct comparison. The changes on stock price were also extracted from the raw transaction data resided in the data warehouse. The values are classified into groups for KDD process. Three algorithms for association rules mining were implemented for efficiency comparison. The mining results are stored in the database server for retrieval through web browser or as input for further processing. The whole system was built using Java with the use of RDBMS and the Java classes built are reusable with defined interface and calling method.
Important Links:
Go Back
