面向功能语义增强与标签关联的Web服务标签推荐

Label recommendation of Web services based on functional semantic enhancement and label association

为了提升标签推荐的质量,提出一种面向功能语义增强与标签关联的Web服务标签推荐方法。将语境权重融入TextRank模型,提取与服务功能契合度高的关键词,用于构建功能语义增强的服务表征向量;建立标签关联图,基于改进的GraphSAGE模型生成标签关联向量;利用KNN算法获取推荐的主标签与候选标签集合,面向服务表征向量和标签关联向量构建融合适配度与关联度的标签推荐方法。实验表明,所提方法在accuracy与F_(1)-score指标上优于当前流行的标签推荐方法,标签推荐质量得到提升。

To improve the quality of label recommendation,this paper proposed a label recommendation method for Web services oriented functional semantic enhancement and label association.It integrated the context weight into the TextRank model to extract keywords that fitted well with the service function,which were used to construct the functional semantic enhanced service representation vector.It established the label association graph and generated the label association vector based on the improved GraphSAGE model.It used the KNN algorithm to obtain the recommended primary label and candidate label set.It used the service representation vector and label association vector to construct a label recommendation method combining fitness and association.Experiments show that the proposed method is superior to the current popular label recommendation methods in terms of accuracy and F_(1)-score,and it improves the quality of tag recommendation.