摘要
提出一种基于最优化控制模型的文本主题域划分方法,采用主题域内距离、主题域间距离、主题域内夹角和主题域间夹角等相关要素,构建了最优化模型的目标函数,进而通过对模型求解得到文本主题域的最优划分模式.该方法为全局最优化方法,与具体的应用领域无关,具有较高的普适性.实验结果表明,在算法适用性、F1评价和Window Diff评价上,该算法均优于其他相关算法.
A subtopic-field segmentation technique based on the optimal control model was proposed, the object function of the optimal control model was constructed by the within-subtopic-field distance, the between- subtopic-field distance, the within-subtopic-field angle and the between-subtopic-field angle. By solving the optimal control model, optimal subtopic-field segmentation is obtained. The method independent of specific application is a global optimal method. The experiments show that this method is better than other methods in applicability, F1 measure and Window Diff measure.
出处
《吉林大学学报（理学版）》
CAS
CSCD
北大核心
2009年第4期769-776,共8页
Journal of Jilin University:Science Edition
基金
国家重点基础研究发展计划973项目基金(批准号:2004CB318000)
关键词
主题域划分
主题域内距离
主题域间距离
主题域内夹角
主题域间夹角
subtopic-field segmentation
within-subtopic-field distance
between-subtopic-field distance
within-subtopic-field angle
between-subtopic-field angle
