维文领域本体组合词概念分类关系提取  

EXTRACTING TAXONOMIC RELATION OF COMBINATION WORD CONCEPTION IN UYGHUR DOMAIN ONTOLOGY

作　　者：杨倩倩[1] 刘胜全[1,2] 刘艳[1] 李连倍 

机构地区：[1]新疆大学信息科学与工程学院,新疆乌鲁木齐830046 [2]新疆大学网络与信息技术中心,新疆乌鲁木齐830046

出　　处：《计算机应用与软件》2015年第7期79-81,101,共4页Computer Applications and Software

摘　　要：针对维文黏着语的特点和广义后缀树提取概念间分类关系时后缀树中出现非概念词的问题,提出一种改进的基于广义后缀树的维文领域本体组合词概念分类关系提取算法。该算法首先对维文领域本体组合词概念构建广义后缀树,先序遍历广义后缀树,对叶子节点存储的后缀词进行维文词干提取,删除非概念词所在叶节点,合并经维文词干提取后表示相同概念的叶节点,实现广义后缀树的剪枝;进而自动提取组合词概念分类关系。实验表明,与传统的基于广义后缀树的概念分类关系提取算法相比,准确率、召回率都得到了提高。According to the characteristics of Uyghur agglutinative language and the problem that the non-conception word appears in the suffix tree at the stage of extracting the taxonomic relation with the method of generalised suffix tree, we proposed an improved generalised suf- fix tree-based extraction algorithm for Uyghur domain ontology combination word conception taxonomic relation. It first constructs the gener- alised suffix tree for Uyghur domain ontology combination word conception, adopts the first traversal generalised suffix tree, extracts the Uy- ghur stemming from suffix word stored in leaf nodes, deletes the leaf nodes where the non-conceptional words are contained, and merges those leaf nodes with same conception after the Uyghur stemming extracted, as well as implements the generalised suffix tree pruning; then it auto- matically extracts the combination word conception taxonomic relation. Experiment showed that the precision and recall rate of this extraction algorithm was increased compared with the conceptual taxonomic relation extraction algorithm based on generalised suffix tree.

关 键 词：维文 广义后缀树 组合词概念 词干提取 分类关系 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]