基于HMM+CRF词性标注的实体抽取方法  

Entity Extraction Method Based on HMM+CRF Part-of-speech Tagging

作　　者：张航 文斌[1] ZHANG Hang;WEN Bin(School of Information Science and Technology,Yunnan Normal University,Kunming 650500)

机构地区：[1]云南师范大学信息学院,昆明650500

出　　处：《计算机与数字工程》2023年第12期2929-2933,共5页Computer & Digital Engineering

摘　　要：基于HMM+CRF词性标注的实体抽取方法从词性标注入手,对待处理文本先进行词性标注;然后根据文本的词性将实体抽取出来,在传统的CRF词性标注模型上增加一层HMM模型,提高实体抽取的精确度;最后在人民日报语料上进行实验,验证了准确率分别在基于HMM实体抽取模型和基于CRF实体抽取模型的基础上提高了2.1%和0.3%。The entity extraction method based on HMM+CRF part-of-speech tagging is injected from the part-of-speech tag.The text to be processed is first tagged with part-of-speech,and then entities are extracted according to the part-of-speech of the text.A layer of HMM model is added to the traditional CRF part-of-speech tagging model to improve the accuracy of entity extrac-tion.Finally,experiments are conducted on the corpus of People's Daily,which verifies that the accuracy rates are increased by 2.1%and 0.3%based on the HMM entity extraction model and the CRF entity extraction model.

关 键 词：实体抽取 隐马尔可夫模型 条件随机场 词性标注 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]