摘要
为了提高领域概念筛选的准确率,对测试数据进行了筛选,指出了算法误筛选常用词汇的问题,分析了常用词汇的高一致度值导致高综合值的缺陷,设计并探讨了术语综合值修正参数,改进了原有综合值的计算方法,放大了领域概念与常用词汇之间的差别。仿真实验表明,修正后的综合值与原综合值相比,变化趋势一致,但幅度更大。数据实验表明,在不影响领域概念筛选结果的基础上,改进后的筛选算法增大了领域概念的综合值,同时降低了常用词汇的综合值,实现了常用词汇的剔除,提高了准确率。
In order to improve the accuracy of domain-specific concept sieving algorithm,the sieving test data was studied, error filtering in some common vocabulary by algorithm were pointed out. The high consensus values of common vocabulary led to high comprehensive value were analyzed, the correction parameters for terminology comprehensive desigued and discussed, the original calculation method improved, and the difference of domain-specific concepts and common vocabulary enlarged. Simulation results show that compared with traditional comprehensive value, the improved comprehensive value has the same variation trend, but larger range.Data experiments results show that the improved algorithm increases the comprehensive value of domain-specific concept, reduces that of common vocabulary,without bringing any impact on the sieving of domain-specific concepts,rejects common vocabulary,and improves the accuracy rate.
出处
《解放军理工大学学报（自然科学版）》
EI
北大核心
2016年第3期203-208,共6页
Journal of PLA University of Science and Technology(Natural Science Edition)
基金
江苏省自然科学基金资助项目(BK2011120)
关键词
综合值
筛选算法
领域概念
comprehensive value
sieving algorithm
domain-specific concept
