The Application Of Data Mining In Stock Analyzing And Predicting

Nowadays, the development of data collection and storage technology makes the data in many databases very large, so it is more and more difficult to get the valuable information and knowledge from these data, however, the requirement of getting this information is growing. The study purpose of data mining is to find the regular information, which is hided in large data set and people are interested in. Data mining is a process of finding knowledge, extracting information, and these knowledge and information can be used for predicting and decision-making.Stock market plays an essential role in economy, effective stock forecast is of great importance in financial investment field, however, stock market is influenced by various complicated factors like policy, economy and investors' mentality etc, so it is a very complicated system with complexity and uncertainty and it is very hard to build a model. Moreover, the regular information is always hided in those massive and disordered stock data, data mining provides a very important method for getting those hidden and valuable information from the massive data. Hence, stock analyzing and predicting has extraordinary theoretic significance and practical value by using data mining.This thesis mainly discusses how to analyze and forecast the stock by using decision tree technology, and a set of classification rules for stock data analyzing and predicting is gotten. The stock investors can predict the stock price based on the rules to reduce the investment risk. The main work is as follows:First, a data mining model has been constructed by considering the characteristics of the stock dealing data. Second, the stock dealing data are pretreated and some analysis indices have been structured. Third, the stock dealing data are mined by using the ID3 algorithm, which is adjusted to meet the needs of this type of problem, a decision tree classifier is generated and a set of classification rules has been gotten from the decision tree, then the results are tested, the testing result shows that this method is feasible and valid. Finally, a stock analyzing and predicting system has been developed according to the above stock dealing data mining model and classification rules, this system can analyze and predict the stock dealing data in real time, and generate some helpful investment information.