摘要
提出一种基于Fuzzy c-means聚类方法的模糊服务聚类方法SGCom,该方法首先标注服务的名称、功能和使用对象,再用改进的FCM算法对服务的名称聚类,用比较相似度的方法对服务的其他元素聚类,并综合得出聚类结果。由于该方法基于服务的注册信息,不局限于单一的服务描述语言,得到的结果是服务属于某个类别的比率,在服务发现时可以根据用户设定的阈值推荐服务类别,避免了靠近类簇边界上的服务难以被准确分类的问题。
In this paper, a Fuzzy c-means based fuzzy service clustering method SGCom is proposed. When using this method, the names, functions, and goals of the Web services are tagged, the improved FCM algorithm is used to cluster the names of services, and similarity of other Web service elements is calculated to cluster them. Comprehensive cluster results can be obtained through these steps. The SGCom method is bases on service registration information and is not limited to a single service description language, so the result obtained is the membership degree of service that belongs to a category, and the category of a service can be recommended according to the threshold set by the user when the service is found, which solves the problem that the services close to the cluster boundaries are difficult to be classified accurately by the hard clustering methods.
出处
《计算机时代》
2017年第11期30-34,38,共6页
Computer Era
基金
华中师范大学2016年校级教学研究项目"基于CDIO风格的面向对象软件工程实践教学模式研究"(201639)
教育部人文社会科学研究规划基金项目"基于学习分析的个体学习者模型构建及服务研究"(15YJA880095)
