python摘要如今NBA是个热门话题,很多专家也会在赛前进行分析,预测当天比赛的情况。那么问题来了:有没有一种好的算法,能够尽量准确地预测出球赛的结果呢?该课题主要是研究找到合适的决策树或者随机森林算法,通过抽取不同的特征,对预测的准确度不断优化,找到预测结果相对准确的算法。在遍历算法的过程中效率不会很高,所以最好一次性处理所有特征,并且每条数据只处理一次。构建决策树时,最后几步决策仅依赖于少数个体,随意性大。使用特定节点作出推测容易导致过拟合训练数据,这时候需要使用退出准则来防止决策精度过高。除了设定退出准则外,也可以先创建一棵完整的树,再对其进行修剪,去掉对整个过程没有提供太多信息的节点。88487
Today the NBA is a hot topic, and many experts will predict the situation of the result before the game。 So the question is: is there a good algorithm that can predict the results of the game as accurately as possible?The subject is mainly to find the appropriate decision tree or random forest algorithm。By exacting the different characteristics, to optimize the accuracy of prediction, and find a relatively accurate algorithm。 The efficiency of the algorithm is not very high, so it is best to deal with all the characteristics, and use each data only one time。 When building a decision tree, the last few steps rely only on a few inpiduals, randomness。 Using a specific node to make inference is easy to lead to fitting training data, which requires the use of exit criteria to prevent high accuracy。 In addition, setting out exit criteria, you can also create a complete tree, trim it, and remove nodes that do not provide much information about the entire process。
毕业论文关键字: 决策树; 随机森林; 预测; 节点
Keyword:decision tree; random forests; predict; node
目录
1 球赛预测及决策树算法的背景介绍 4
1。1 球赛预测简介 4
1。2 现有的预测方法 4
1。3 决策树算法背景介绍 5
1。4 python背景源Y于Y优E尔Y论L文W网wwW.yOueRw.com 原文+QQ752018.766 介绍 5
2 决策树及随机森林 6
2。1 决策树算法 6
2。2 随机森林算法 8
2。3 相关库、函数等介绍 9
3 对nba2015赛季做实例计算 10
3。1 研究的总体概述 10
3。2 pandas加载库 11
3。3 决策树算法 13
3。4 随机森林算法 15
4 总结 16
参考文献 17
致谢 17
1 球赛预测及决策树算法的背景介绍 python基于决策树算法的球赛预测:http://www.youerw.com/jisuanji/lunwen_162241.html