摘要
针对军事领域的知识图谱的应用,设计实现了基于军事领域知识图谱的合成旅指挥员智能问答系统。首先,通过对军事领域应用的特点分析,设计实现了基于Jieba工具的中文分词模块、基于双向长短期记忆网络-条件随机场(Bidirectional Long Short Term Memory-Conditional Random Field,BiLSTM-CRF)的命名实体识别模块和基于Cypher语言的问题查询模块等关键模块;然后,构建了基于环球军事网以及新浪军事中爬取到的新闻数据构建数据集,进而基于此数据集进行命名实体识别算法实验对比分析;最后,对BiLSTM-CRF算法进行参数调优,使得模型的识别效果达到最优,进而对系统进行了展示。
In this paper,aiming at the application of knowledge mapping in the military field,an intelligent question-and-answer system for commanders of combined brigades is designed and implemented.Firstly,by analyzing the characteristics of application in the military field,key modules such as Chinese word segmentation module based on Jieba tool,named entity identification module based on Bidirectional Long Short Term Memory-Conditional Random Field(BiLSTM-CRF)and question query module based on Cypher language are designed and implemented.Then,a data set based on the news data crawled from the Global military network and Sina military is built,and on this basis,,the experimental comparative analysis of named entity recognition algorithm is carried out.Finally,the parameters of BiLSTM-CRF algorithm are optimized to achieve the best recognition effect of the model,and then,the system is demonstrated.
作者
王宏宇
许潇
周育伟
杨朝红
纪伯公
WANG Hong-yu;XU Xiao;ZHOU Yu-wei;YANG Chao-hong;JI Bo-gong(Army Academy of Armored Forces,Beijing 100072,China;Department of Computer Science and Technology,Tsinghua University,Beijing 100089,China;Army Staff Department,Beijing 100042,China)
出处
《装甲兵学报》
2022年第2期87-94,102,共9页
Journal of Armored Forces
