基于语法语义知识的维吾尔文机构名识别  

Uyghur organization name recognition based on syntactic and semantic knowledge

作　　者：麦合甫热提[1] 米日姑.肉孜 麦热哈巴.艾力[3] 吐尔根.依布拉音 

机构地区：[1]新疆大学教务处,新疆乌鲁木齐830046 [2]新疆大学多语种信息技术重点实验室,新疆乌鲁木齐830046 [3]新疆大学信息科学与工程学院,新疆乌鲁木齐830046

出　　处：《计算机工程与设计》2014年第8期2944-2948,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61262061;61262060;61063026);国家社科基金重点项目(10AYY006)

摘　　要：为了提高维吾尔语中机构名的自动识别准确率,从维吾尔语的语言特点出发,对维吾尔语中机构名的组织结构进行了分类并将其形式化表示;根据此特征设计出有效地识别规则,创建了特征词库、地名库和修饰词库等知识库;设计并实现了基于状态转移原理的高效识别算法。实验结果表明,该算法识别的F值达到83.05%,获得了较好结果。To improve the automatic recognation of organization name in Uyghur, through anaiyms of the charactersitics of Uyghur organization name, the following work was done. First, the organization name in Uyghur was classified depending on its structure and it was formally described. After then, effective recognizing rules were desingned according to these features, knowledge base was created such as features word base, place name base and qualifier word base. Finally, efficient recognition algorithm was designed and implemented based on the principles of state transition. Representative examples from the Tianshan net news were selected to build the test set for organization name recognition, experimental results showed that, this method achieved better results with the F measure of 83.05 %.

关 键 词：自然语言处理 命名实体识别 机构名识别 知识库 规则匹配 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]