Self-organizing Data Mining For Applied Research In The Stock Market

With the rapid development of information technology, the stock market has cumulated a flood of data. So how to discovery knowledge from the data has become the hot topic.As an often-used data mining tool, statistical model still has some disadvantages. One of the most significant disadvantages is that some rigorous hypotheses are often indispensable. Hence, they are frequently failed to deal with some realistic problems. This dissertation introduces a different data mining tool, Self-Organising Data Mining (SODM). Self-Organising Data Mining bases on the principles of evolution-inheritance, mutation and selection-for generating a network structure systematically enabling automatic model structure synthesis and model validation. This dissertation explains the idea of Self-Organising Data Mining, introduces the algorithm of Self-Organizing Data Mining, and describes how Self-Organizing Data Mining works.According to the different features of two kinds of data in stock market, we use GMDH and AC to extract knowledge respectively.We use GMDH to set up a forecasting model of Chinese listed companies'financial crisis by using financial data. We compare the prognosticating effects of 3 models, which are set up according to the following 3 kinds of ways: the Discriminate analysis, the Logit analysis, and find that the GMDH is more effective way to transform data into knowledge.We use AC to get some predictive patterns from share price, and then use these patterns to forecast the share price, and simulation results indicate this mining method is valid.This dissertation provides an introduction about Self-Organizing Data Mining technology for the investor in the stock market, and we hope our research can help them to improve the decision making based on analyzing data.