基于知识地图拓扑的核心知识单元识别方法  

An Identification Method of Core Knowledge Units with Topology of Knowledge Map

作　　者：何绯娟[1] 缪相林[1] 许大炜[1] 

机构地区：[1]西安交通大学城市学院,陕西西安710018

出　　处：《计算机技术与发展》2017年第7期34-37,42,共5页Computer Technology and Development

基　　金：教育部人文社会科学研究青年基金项目(15YJCZH057);陕西省社会科学基金项目(2016N005);陕西省科学技术研究发展计划工业科技攻关项目(2015GY012);西安交通大学城市学院校级科研项目(2016KZ01;2015KZ09)

摘　　要：核心知识单元的识别有助于引导学习者的注意力分配,消除学习迷航问题。知识单元之间在认知上具有长距依赖性,常用的度、紧密性、介数、特征向量等中心度指标很难适用于识别知识地图中的核心知识单元。为此,提出了一种基于知识地图拓扑的核心知识单元识别方法。该方法依据对知识地图拓扑分析发现的三个特性,即核心知识单元的层次分布特性、出度分布特性、Motif结构特性,建立了知识单元对应的六维特征向量。在六维特征向量的基础上,将核心知识单元识别问题转化为二类分类问题,采用分类算法实现核心知识单元的识别。在8门课程知识地图数据集上,采用支持向量机(SVM)、决策树C4.5、朴素贝叶斯(NB)和多层感知器(MLP)四种算法进行了对比实验。实验结果表明,所提出的方法有效可行。The recognition of core knowledge units is helpful for guiding the learners＇ attention allocation, eliminating disorientation prob- lem. Because these are long-distance dependences on cognition between knowledge units, it is difficult to identify the core knowledge u- nits from a knowledge map by using the traditional centrality indexes, such as degree, closeness, betweenness and eigenvector. An identifi- cation method of core knowledge units based on knowledge may topology is proposed, which establishes six-dimensional feature vectors corresponding to knowledge units according to three characteristics like hierarchical distribution, out-degree distribution and Motif struc- ture through the topological analysis of knowledge map. On the basis of six-dimensional feature vectors, the core knowledge unit identifi- cation is transformed into a binary classification problem, and a recognition method is implemented by using classification algorithm. With knowledge map dataset of eight courses, a comparative experiment has been conducted with four algorithms, including Support Vector Machine （SVM）, Decision Tree （C4.5）, Nalve Bayes （NB） and Multi-Layer Perceptron （MLP） and its result demonstrates the effec- tiveness of the proposed method.

关 键 词：知识地图 拓扑分析 知识单元 MOTIF 中心度 

分 类 号：TP391.2[自动化与计算机技术—计算机应用技术]