基于层次模型的科学工作流聚类和推荐

发布时间：2018-04-22 17:01

本文选题：层次模型 + 科学工作流　；参考：《中国地质大学(北京)》2017年硕士论文

【摘要】：随着Web技术的发展,分布在因特网中不同节点处的应用程序包装为Web服务,并突破平台结构的差异相互调用。为了构建具有新功能的服务,对已有服务进行组合,形成复合服务链。随着服务组合推荐技术的发展,常用的复合服务作为科学工作流进行共享。科学工作流是由活动(服务)、活动(服务)之间的依赖关系以及子工作流组成。科学工作流的急速增长促进了其发现、共享和重用。当科学工作者构建新的功能实验时,重构新的工作流需要丰富的领域知识,以及对科学工作流中活动以及活动之间的依赖关系的深入了解。在此情景下,科学工作流的重用成为重要研究课题。目前,国内外研究者对此问题进行了广泛而深入地研究。其中,如何对科学工作流的适配性进行评估是当前研究热点之一。传统工作通过科学工作流的结构或语义标识的途径对科学工作流的适配性进行探讨。然而,这些工作对工作流、子工作流和活动之间层次性关系的研究尚不足,即对多粒度科学工作流的适配性的探索有待加强。为解决该问题,本文提出了基于层次模型的科学工作流重用算法。具体内容如下:1)通过对科学工作流结构的分析,针对科学工作流与活动、子工作流之间的关系,提出通过层次性模型表示科学工作流在多粒度下的功能性描述,从而表征科学工作流的结构以及语义性信息;2)基于树的编辑距离和语义相似度算法,提出基于层次模型的科学工作流适配性评估方法,考虑了科学工作流的层次性和工作流树状结构以及语义信息。从而构建科学工作流网络模型,表示科学工作流之间的相似性关系;3)基于科学工作流网络模型,提出基于图骨架聚类算法对科学工作流进行聚类,并发现核心工作流,代表相应聚类。提出两个阶段的科学工作流的推荐算法。基于初始阶段确定的目标聚类,算法在第二阶段可快速检索符合用户需求的科学工作流,并进行排序和推荐。基于本文提出的科学工作流聚类以及推荐算法,通过使用my Experiment科学工作流数据集,应用不同聚类算法以及实验评估函数,对比验证了本文聚类算法的有效性与准确性。
[Abstract]:With the development of Web technology, applications distributed at different nodes in the Internet are packaged as Web services, and the differences in platform structure are called to each other. In order to construct new services, the existing services are combined to form a composite service chain. With the development of service composition recommendation technology, common composite services are shared as scientific workflow. Scientific workflow is composed of activities (services), dependencies between activities (services) and sub-workflows. The rapid growth of scientific workflow promotes its discovery, sharing and reuse. When scientists construct new functional experiments, reconstructing new workflows requires rich domain knowledge and in-depth understanding of activities and their dependencies in scientific workflows. In this situation, the reuse of scientific workflow has become an important research topic. At present, researchers at home and abroad have carried out extensive and in-depth research on this issue. How to evaluate the adaptability of scientific workflow is one of the current research hotspots. The traditional work discusses the adaptation of scientific workflow by the way of structure or semantic identification of scientific workflow. However, the research on the hierarchical relationship among workflow, sub-workflow and activity is not enough, that is, the research on the adaptability of multi-granularity scientific workflow needs to be strengthened. In order to solve this problem, a hierarchical model based reuse algorithm for scientific workflow is proposed. The specific contents are as follows: 1) based on the analysis of the structure of scientific workflow and the relationship between scientific workflow, activity and sub-workflow, a hierarchical model is proposed to represent the functional description of scientific workflow under multi-granularity. Therefore, the structure of scientific workflow and semantic information are represented. Based on tree editing distance and semantic similarity algorithm, a hierarchical model based assessment method for scientific workflow adaptability is proposed. The hierarchy of scientific workflow, workflow tree structure and semantic information are considered. Based on the scientific workflow network model, a graph skeleton clustering algorithm is proposed to cluster the scientific workflow, and the core workflow is found. Represents the corresponding clustering. A two-stage scientific workflow recommendation algorithm is proposed. Based on the target clustering determined in the initial stage, the algorithm can quickly retrieve the scientific workflow according to the user's needs in the second stage, and make sorting and recommendation. Based on the scientific workflow clustering and recommendation algorithm proposed in this paper, by using my Experiment scientific workflow data set, using different clustering algorithms and experimental evaluation functions, the validity and accuracy of this clustering algorithm are compared and verified.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3;TP393.09

【相似文献】