改进的C4.5算法在CET-4成绩分析中的应用  

Research on the application of the improved C4.5 algorithm in the analysis of CET-4 scores

作　　者：王善勤 王立辉[1] WANG Shanqin;WANG Lihui(School of Instrument Science and Engineering,Southeast University,Nanjing 210096,China;School of Information Engineering,Chuzhou Polytechnic,Chuzhou 239000,China)

机构地区：[1]东南大学仪器科学与工程学院,南京210096 [2]滁州职业技术学院信息工程学院,安徽滁州239000

出　　处：《黑龙江工程学院学报》2023年第2期34-40,共7页Journal of Heilongjiang Institute of Technology

基　　金：安徽省高校优秀拔尖人才培育资助项目(gxgnfx2020159);安徽高校自然科学研究重点项目(KJ2021A1408)。

摘　　要：CET-4是一个客观、准确的大学生英语能力测量平台,C4.5算法在应用于CET-4成绩分析中仍存在一些问题。针对运用C4.5算法对高职院校CET-4成绩数据构建分析决策树时存在的离散化运算繁琐、忽视各属性影响度等典型问题,提出一种面向高职院校CET-4成绩分析的改进C4.5算法。首先通过在C4.5算法中引入成绩正态分布规律确立初始聚类中心、K-means算法来离散连续属性;其次引入CET-4中听、读、写的权重来修正信息增益率的计算;最后运用改进的C4.5算法、经典的C4.5算法分别构建决策树模型并进行预测分析。实验结果表明,改进的C4.5算法所构建高职院校CET-4成绩分析的模型效率、预测能力均有明显提高。运用改进的C4.5算法有效地分析出影响CET-4达标各因素间的关系,从而提升CET-4反拨英语教学效应。CET-4 is an objective and accurate platform for measuring college students’English ability.There are still some problems in the application of the C4.5 algorithm in CET-4 score analysis.In view of the typical problem such as cumbersome discrete operation and ignoring influence of attributes when using C4.5 to construct analysis decision tree for the CET-4 score data of higher vocational colleges,an improved C4.5 algorithm is proposed in this paper.Firstly,the normal distribution rule is used into the algorithm to establish the initial clustering center and K-means algorithm to discrete continuous attributes.Secondly,the weights of listening,reading and writing in CET-4 are introduced to correct the calculation of information gain rate.Finally,the improved C4.5 and Classic C4.5 algorithm are used to construct the decision tree model and make a comparative analysis of prediction.The experimental results show that the improved C4.5 algorithm has significantly improved the efficiency and prediction ability of constructing the CET-4 score analysis model in higher vocational colleges.The improved C4.5 algorithm can be used to effectively analyze the relationship between the factors affecting the achievement of CET-4,so as to improve the effect of CET-4 backwash English teaching.

关 键 词：CET-4 正态分布 K-MEANS C4.5算法 决策树 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]