迁移知识辅助的语义稀疏服务聚类方法

中文摘要: 现有服务聚类方法缺乏对服务描述语义稀疏情境下的研究，因此将迁移学习技术应用到服务聚类领域，尝试解决语义稀疏服务聚类的问题。通过对偶PLSA模型将目标领域和辅助领域语料知识进行融合，利用监督的方式迁移辅助领域知识,从而提高目标领域语义稀疏服务聚类的能力。实验结果表明，该方法能够提高语义稀疏服务的聚类效果。与 K-Means、Agglomerative和PLSA等方法相比，该方法在聚类纯度、熵上均具有更好的性能。
Abstract:The existing clustering approaches are lacking of researching on clustering Web services whose descriptions are semantic sparse.Therefore,a new approach was proposed to makes an attempt to solve the problem of clustering sematic sparse Web service by applying transfer learning method in the domain of Web service clustering.A dual PLSA model was introduced to integrate knowledge of target domain and auxiliary domain which can transfer knowledge using a unsupervised mode to facilitate the process of semantic sparse Web service clustering.Experimental results showed that the proposed method can improve the performance of semantic sparse Web service clustering.Compared with the approaches of K-means,Agglomerative and PLSA,the proposed approach achieves better performance of the purity and the entropy.