一种基于邻域粗糙集的多标记加权分类算法

A MULTI- LABEL WEIGHTED CLASSIFICATION ALGORITHM BASED ON NEIGHBORHOOD ROUGH SETS

多标记分类问题在文本分类、图像标注和基因功能组学习等领域都有很好的应用前景．考虑到多标记分类问题中的不确定性和相关性问题，引入了邻域粗糙集模型来构造一种新的框架MLRS．但该方法忽略了邻域中样例与测试样例之间的局部相关性，针对该问题，提出一种新的加权多标记分类算法WMLRS．实验表明，本文提出的方法拥有比其他常用的多标记分类算法更优的分类性能．

Multi -label classification problems have a good application prospect in the areas such as text classification, image annotation and gene functional group learning, etc. Neighborhood rough set model can be introduced to construct a new framework called MLRS by considering the uncertainty and correlation questions in multi -label classification problems. But this algorithm ignored the local correlation between the labels of the test examples and their k - nearest neighbors. To cover this defect, this paper proposed a new algorithm called WMLRS algorithm. Experimental results show that the proposed approach WMLRS works better than other commonly used multi - label algorithms.