Data Mining Of Fencing Athletes' Blood Items Optimization

To improve match grade, athletes must have rational, systemic, scientific training. In recent years, scientific training methods have being more and more used in sports and have made distinct effect. In numerous sports, fencing is important sport for Chinese team to win medals in ATHENS 2004 and BeiJing 2008 Olympic Games. To help coaches of fencing team coach more scientifically and help athletes improve match grade, we researched a software named "Fencing Training Burthen Computer Analysis System". An important research content of this system is using data mining and knowledge discovery to evaluate athletes' status and sport burthen and help coaches make training plan more rationally.Now, based on testing record of blood items, we have implemented the evaluation of athletes' physical status using Neural Networks. However, we often choose blood items randomly, and this will influence the performance and precision of Neural Networks. Therefore, we must choose rational blood items before training Neural Networks. The objective of this paper is how to optimize blood items. We found the relations of blood items and implement the optimization of blood items using association rules which is a very important method of data mining.The form of association rules is X (?) Y, and the meaning of which is if one record contains X, it is likely to contain Y. The discovery of association rules have two steps. The first step is to find all frequent itemsets which needs the support of frequent itemsets is not less than the minimum support made by users. The second step is to create rules from frequent itemsets, and the confidence of rules is not less than the minimum confidence made by users. To find all frequent itemsets is core of algorithms of association rules and is the most complex part of computation. The most famous algorithms of association rules is Apriori put forward by R.Agrawal. The algorithm uses an iterative method which use k-itemset to find (k+l)-itemset. Algorithm of Apriori may produce a mass of candidate itemsets and scan databaserepeatedly which will influent the efficiency of algorithm. The other often used algorithm of association rules is Frequent Paten Growth which do not produce candidate itemsets and improve the efficiency a lot.Algorithms of association rules are based on transaction database. Because we save data using relation database SQL Server, we have to ameliorate two algorithms of association separately named Apriori and FP-Grwoth. We gave the implement methods of this two algorithms in relation database in the end.This paper first researches data mining and pick-up association rules, secondly gives the implement methods of Apriori and FP-Growth in relation database, finally gives the introduction of the application of optimization of blood items.