Research And Application Of Data Mining In Stock Analysis

Data mining is also called Knowledge Discovery in Databases, it is from the datasource (such as text, pictures, Web, database, etc.) to find useful patterns in the process.These models must have the potential practical value, and can be understood and applied.Data mining are not only to master the advanced mining technology, and moreimportant to have a relevant business knowledge and thinking. Possess a wealth ofbusiness knowledge related to the design of effective models and variables, so as tocombine the technical and business efficient access to valuable data.Factors that affect stock price changes a lot, so the use of some fixed mathematicalformula is difficult to accurately describe the trends in the stock. In a variety of stockpredicting methods, the use of association rules to predict the ups and downs of the stockhas a guiding value. This project through in-depth knowledge of stock research, stock datacollection, cleaning, extraction, modeling, analysis of single stock over the years to themonth and a week of ups and downs over the years a situation, look at the changing stateof the same the state. On multi-year stocks do correlation analysis, to find its ups anddowns associated with the probability between states, calculated within three consecutivedays of stock in the association rules as well as single stocks of N days and M days by thestate directly study the relationship between contrast and calculate their support andconfidence in order to predict the ups and downs of the stock price analysis. Thecharacteristics of the stock data, stock data using the bitmap model, using the apriorialgorithm for mining frequent itemsets, computing support and confidence, to findfrequent item sets in when using the bitmap and the KMP matching algorithm to computethe common accelerate the matching speed.