Modeling And Algorithm Analysis Stock Market Index Based On Association Rules

Large amounts of data exist in the stock market, investors need to analyze these data and make use of them. Data mining brings a new method for the analysis of stock market. The association rules mining can find the interesting connection or the related relation among the item sets in mass data, it is an important area in data mining. There is a remarkable plate linkage phenomenon in the stock market, that is, the stocks in a fundamental category or with a certain concept rise and fall together. This paper based on association rules mining on the stock market plate indexes modeling analysis and algorithm research, it analyses price trends in the stock market and industrial relations, it can provide certain help for investors undertake portfolio and make reasonable decision.This paper introduced the background knowledge of data mining and the stock market, analyzed the meaning of data mining in the analysis of the stock market, and summarized the current application of data mining in the stock market mining, pointed out the value of the stock plate analysis for investors in the stock market. Classic algorithm of association rules had the deficiency of frequent scanning database and producing a large number of candidate items, in view of the deficiency of Apriori algorithm some improvements were made, the improved algorithm using vertical data format of matrix, and added a flag bit, the connection method of generating candidate items was changed too. Then the improved algorithm was applied to the stock plate index analysis, to explore the relationship among the stock plates, and the results were analyzed, then we pointed out how to select valuable rules. Temporal association rules could find the set of temporal relation, and the relation could be used to analyze price trend of the relationship between the plates, in this paper, the classical scheduling algorithm E-Apriori and EH-Apriori were researched. With the combination of the time window method and the improvement of Apriori algorithm, we proposed an improved temporal association algorithm. The new algorithm was applied to the stock plate indexes analysis, and the results were analyzed, the results had certain value on investment analysis. Finally, this paper built a prototype system to verify the above algorithms, it allowed the user to set parameters and display the data and results, and the usefulness of the above algorithms was verified.