基于PR模型的教育数据矩阵加权正负模式研究  

Research on the matrix-weighted positive and negative patterns from educational informatization data based on PR model

作　　者：余如[1] 冯平[2] 黄名选[3] 

机构地区：[1]广西教育学院党政办,南宁530023 [2]广西科技大学电气与信息工程学院,广西柳州545006 [3]广西教育学院科研处,南宁530023

出　　处：《信息技术》2014年第6期9-13,共5页Information Technology

基　　金：国家自然科学基金项目(61262028);广西自然科学基金项目(2012GXNSFAA053235);广西教育厅科研项目(2013LX236;201203YB225);广西高校优秀人才资助计划项目(桂教人[2011]40号)

摘　　要：提出一种新的基于PR(Probability Ratio,简称PR)模型的教育信息化数据矩阵加权正负关联模式挖掘算法。算法从教育信息化数据中挖掘矩阵加权频繁项集和负项集,采用PR模型作为正负关联模式评价标准,对频繁项集和负项集挖掘教育数据矩阵加权正负关联模式,发现有趣的教育信息矩阵加权正负关联模式。通过模式分析,发现教育领域信息化数据中潜在的、有用的教育信息、教学规律和教育发展趋势。以真实的教育信息化数据作为实验数据,实验结果表明,该算法更有效、合理。This paper presents a new mining algorithm of the matrix-weighted positive and negative association patterns from educational information data based on PR （Probability Ratio, referred to as PR） model. The algorithm mines the matrix-weighted frequency itemsets and negative itemsets from educational informatization data, and uses the PR model to measure the matrix-weighted positive and negative association patterns so as to find interesting matrix-weighted positive and negative association patterns. Analysis of these patterns shows that, the potential educational and teaching rules, as well as educational development trend are discovered. It displayed from the experiment on real educational information data that, compared with the existing association patterns discovery algorithm, the algorithm proposed in this paper is more effective and more reasonable.

关 键 词：教育信息化 关联模式 PR模型 矩阵加权模式 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]