Research On Association Rules Mining Algorithm And Its Application In Securities Trading

The technologies of computers, network and the big data are widely used in China financial markets, especially in the stock market. As stock marketsis a representation of a country%s economic, it is significant meaning to research on stock markets and stock trading for China%s economics and politics.In thedata mining process of stock trading, data mining plays a very important role. In essence, data mining refers to the process of searing hiding information from the huge trading data based on some mining algorithms. The mining of association rules, which has been used globally in stock market trading analysis, is a very important aspect of data mining. Basically speaking, the mining of the mining of association rules is the process that can reveal the hidden dependencies between a large numbers of data. According to the relationship that it can be inferred from the relevant object information, investors can find the relationship between stock trend and stock data, so they can make correct invest decision, the mining of association rules in stock market has significant meaning. This paper focus on theory, technology and practice of the mining of the mining of association rules in stock exchange, the main works are as following.Firstly, this paper summarizes data mining and the mining of association rules from theory, technology, problems need to solve, applications in real life aspects, and then discuss application and future trends of the mining of association rules in stock markets especially in stock trading.Secondly, an improved Apriori miningalgorithm of association rules based on subtract searching items and weight parameters has been proposed. After deeply research on the mining of association rules algorithms, we analysis traditional Apriori algorithm, considering characteristics of stock trading time series data, we overcome the disadvantages of Apriori algorithm and proposed an improved mining algorithm of association rules. The new algorithm focus on optimizing steps, processes and parameters of the traditional Apriori algorithm while it has been used in new type stock trading and stock index trading. At last, we analysis the effectiveness and correctness of new algorithm in theory.Finally, a prototype system based on the improved Apriori mining algorithm of association rules, which can be used in stock trading analysis and forecasting, has been designed and implemented in this paper. In the prototype system, we have considered characteristics of China stock market, and convert last five years% stock trading data in to long term parameters of stock market by time series matching method in the mining of association rules, convert last five days, ten days and thirty days% stock trading data into short term parameters of stock market index. After this, the complete parameters collection has been setup which play as the input of the verification program. The prototype system has been implemented using Microsoft Visual Studio, and the effectiveness and correctness of the new algorithm has been verified which provide the new algorithm works well in new type stock trading and stock index trading.