欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
在文本分类中,传统单标签分类问题的解决方法无法简单地应用于多标签文本分类,现有的方法通常会通过单标签问题转化思想或者多标签自身算法改进实现对多标签的文本分类。提出一种相关信息加权的自适应多标签分类算法,该算法具有相关信息加权、自适应阈值调整、权重投票相结合的特点。实验结果表明,该算法的某些性能指标优于现有一些常用的多标签分类方法。
英文摘要：
In text classification area, the solution of traditional methods for single-label classification cannot be simply applied to multi- label text classification, and current methods usually implement the text classification of multi-label by improving the single-label problem transformation idea or the multi-label algorithm its own. What is put forward in this paper is an adaptive multi-label classification algorithm with related information weighting, it features in the combination of related information weighting, adaptive threshold adjustment and voting by weight. Experimental results show that some performance indexes of the algorithm are superior to current common multi-label classification methods.
同期刊论文项目
同项目期刊论文
期刊信息
位置：