基于属性相关的智能学习指导模型的设计与实现  

Design and implementation of intelligent learning guidance model based on attributes correlation

作　　者：张春飞[1] 李万龙[2] 魏久鸿[1] 

机构地区：[1]吉林大学计算机科学与技术学院,吉林长春130062 [2]长春工业大学计算机科学与工程学院,吉林长春130012

出　　处：《河北科技大学学报》2012年第6期525-529,共5页Journal of Hebei University of Science and Technology

基　　金：吉林省教育厅"十二五"科学技术研究项目(2011101)

摘　　要：决策树分类算法是智能指导系统实现"智能"的一种有效工具。通过对数据的分析和挖掘,能够实现对数据的精确分类。另外,对于决策树和产生式规则集的计算相对简单而且高效。提出了智能指导系统,并介绍了该系统的主要功能模块。在比较了ID3算法和C4.5算法后,结合个性化教学的需求,提出了新的基于规则属性相关的C4.5r算法。同时,给出了系统的计算评估模块。实验结果表明,新的C4.5r算法在运算时间、产生式规则集的规模及计算产生式规则的开销方面明显优于传统的C4.5算法。Decision tree classification algorithm is one of the effective tools to realize ＂intelligent＂ of Intelligent Guidance Sys- tem. It can receive precise classification through the analyzing and mining of the data. It also has another positive characteristic that the decision tree and the set of production rules are simple and efficient in terms of computing. This paper proposed the ＂Intelligent Guidance System＂ and introduced the main modules. It also proposed C4.5r algorithm based on the comparison of ID3 algorithm and C4.5 algorithm and the requirements of individual educations. Simultaneously, the model used for evaluation was established. Experiments show that C4.5r algorithm is better than C4.5 algorithm in such aspects as run-time, the size of rules sets and overhead of the production rules.

关 键 词：智能学习 WEB数据挖掘 决策树 建构主义理论 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]