基于知识库的自然语言中歧义字段自动识别系统设计  

Design of Automatic Recognition System for Ambiguous Fields in Natural Language Based on Knowledge Base

作　　者：卫欣玲[1] WEI Xin-ling(Shanxi college of communication technology,Xi'an 710018 China)

机构地区：[1]陕西交通职业技术学院,陕西西安710018

出　　处：《自动化技术与应用》2023年第1期69-72,151,共5页Techniques of Automation and Applications

摘　　要：为全面实现歧义消解,对知识库自然语言中歧义字段自动识别系统设计。利用互信息进行歧义字段特征识别,采用正向与逆向相结合的提取方式,将字段特征集合描述为二维向量,通过循环方式提取歧义字段显性特征;建立模型进行最优线性分类识别,选择最佳样本识别条件,建立最优分类超平面并确立分类函数。通过性能衡量指标构建软硬件系统结构,结合识别算法设置工作流程,经功能设计进一步提高识别精度。仿真实验表明该系统不受数据规模影响,可有效提高识别精度,减少系统处理时间,实现对歧义字段的高精度、高效率识别。In order to realize ambiguity resolution, the automatic recognition system of ambiguous fields in natural language of knowledge base is designed. Mutual information is used to identify the features of ambiguous fields. The feature set of fields is described as two-dimensional vector, and the dominant features of ambiguous fields are extracted in a cyclic way. The optimal linear classification and recognition conditions are selected, and the optimal classification hyperplane and classification function are established.The software and hardware system structure is constructed by performance measurement index, and the workflow is set up with recognition algorithm. The recognition accuracy is further improved by functional design. The simulation results show that the system is not affected by the data scale, and can effectively improve the recognition accuracy, reducing the processing time of the system, and realize the high-precision and high-efficiency recognition of ambiguous fields.

关 键 词：文字自动识别 支持向量机 

分 类 号：TP391.11[自动化与计算机技术—计算机应用技术]