赛事分析与预测的KDD建模

把KDD用于赛事决策支持，见之报道的还很少，本文针对一类非对垒式的，以个体成绩为主进行名次排序的体育和娱乐赛事，给出了一个该类赛事分析与预测的KDD建模方法，用于帮助分析和确定对每个参赛者的比赛成绩有影响的各种重要因素，并对比赛结果作出预测。我们的方法以轻量的、多模型的和多种技术的组合／结合为策略；兼顾到了问题的各个方面和主要特征：可以透过对几个参数的简单处理，自动、方便地对已建立的模型进行不断的修正；同时还能主动建议对模型的改进或重建，并为模型的改进或重建提供有用的帮助信息，从而能够较好地解决参赛群体多样性、多变性、影响因素的复杂性和预测的困难性等问题。

A MODELING METHOD FOR GAMES ANALYSIS AND PREDICTION BY USING KDD TECHNIQUES

Few formal details have been reported about using KDD to support decision-making in games.For a kind of non-confrontation games of sports and entertainments in which the names of contestants are arranged in the order of their results,we present in this paper a modeling method for the analysis and prediction of that kind of games by using KDD techniques.The method,aiming to assist analyzing and determining the significant factors that affect the scores of each contestant and making prediction about the game's final results,takes the strategy of composition/combination of lightweight models and techniques;takes every aspect and primary characteristics of the problem into account.It can make continuous revision to the built model automatically and conveniently,where only some simple adjustments of several parameters are needed.Besides,it can also make active suggestion about when to improve or rebuild the model and provide useful information to help the improvement and rebuilding.Therefore the method provides a fairly good solution to the problem,which is noted for the variety and volatility of contestants,complexity of the factors and the difficulty of prediction.