The Research And Implementation Of Ice Hockey Coach Assistant System

With the successful bid of the 2022 Winter Olympics,China's ice hockey has a new development.The daily training and competition of hockey players will generate a lot of information,how to find hidden content in the complicated information,and how to make science and reasonable teaching plans are the particular concern for coaches at all levels.This paper designs a hockey instructor assisted system based on data mining,and focuses on the key techniques of local outliers identification and the analysis of athletes outliers,in order to meet the needs of the system for efficient and scientific analysis the hockey player's performance.The main research contents of this paper are as follows:(1)The requirements of the ice hockey coaching aid system are analyzed.The overall framework and system function design of the system are given.The function design of each module of the system is described in detail,and the key technologies involved in the system are introduced in detail.(2)An outlier athlete recognition algorithm based on improved LOF algorithm is studied.The algorithm is divided into two stages: obtaining preliminary outlier data sets and outlier data identification.The algorithm firstly preprocesses the original data set by using the clustering algorithm OPTICS to obtain a preliminary outlier data set.When calculating the distance between data objects,the weights of different attribute items are determined by information entropy increment.Because the P weight can only detect the outlier data of a single density,the P weight is used to replace the reachable distance in the LOF algorithm,which improves the accuracy of the algorithm outlier data detection.The algorithm is verified by experiments.The results show that the algorithm can improve the efficiency of calculating local outliers and the accuracy of outliers.(3)The outlier cause analysis algorithm based on genetic algorithm and fuzzy frequent itemsets is studied.Since the performance attribute value of the athlete is quantitative,and the traditional frequent item set mining algorithm can only deal with the boolean attribute,the method of obscuring the quantity attribute by the membership function in the fuzzy set is given.In order to improve the deficiencies of the membership function determined by the manual,the algorithm uses the genetic algorithm to generate the membership function.And in order to make the division of the fuzzy interval more reasonable,the algorithm redefines the fitness function and optimizes the membership function according to the hockey performance characteristics.When mining frequent itemsets,the algorithm proposes a new strategy to filter fuzzy attribute regions by introducing variable thresholds,in order to preserve more fuzzy regions.The experiment proves that the algorithm can generate the membership function which is more reasonable and more suitable for hockey score analysis,and the mining results contain more fuzzy regions and better mining ability of fuzzy frequent itemsets.(4)Developed and implemented the ice hockey coaching assistant system,which mainly achieved the following functions: athlete ice hockey game score entry and management,coach information management,access authority management,outlier athlete identification,analysis of the reasons for athletes' outliers,data analysis results visualization,tactical board design and other system main functions,giving the operating interface of each function.