| 引用本文: | 常关羽,杨海成,莫蓉,孙鹏.面向综合特征提取的流程相似度模型[J].哈尔滨工业大学学报,2017,49(7):183.DOI:10.11918/j.issn.0367-6234.201604062 |
| CHANG Guanyu,YANG Haicheng,MO Rong,SUN Peng.Feature extraction oriented similarity metric of business process[J].Journal of Harbin Institute of Technology,2017,49(7):183.DOI:10.11918/j.issn.0367-6234.201604062 |
|
| 摘要: |
| 针对流程相似度计算研究中注重流程结构而缺乏兼顾流程语义的问题，以及现有相似度计算方法在计算复杂度上的不足，提出一种基于流程综合特征提取的相似度计算模型.基于流程基本控制结构分析，提出边权重标注方法以扩展现有流程结构，提取流程结构特征；定义流程高层语义模型及其对应特征提取方法；融合了节点集、边集相似度，给出新的流程结构相似度定义，利用集合关系和向量空间模型计算流程语义相似度；通过加权实现综合流程相似度评价，并采用权重参数调节的方式实现了同已有相似度计算方法的自适应转化.将本文模型与典型相似度计算方法进行了实验对比，结果表明, 面向综合特征提取的流程相似度计算方法更具普适性，同时具有更高效的计算能力.
|
| 关键词: 流程模型 业务语义 特征提取 流程相似度 流程匹配 |
| DOI：10.11918/j.issn.0367-6234.201604062 |
| 分类号:TP315 |
| 文献标识码:A |
| 基金项目:国家自然科学基金(51375395) |
|
| Feature extraction oriented similarity metric of business process |
|
CHANG Guanyu,YANG Haicheng,MO Rong,SUN Peng
|
|
(Key Laboratory of Contemporary Design and Integrated Manufacturing Technology, Ministry of Education (Northwestern Polytechnical University), Xi’an 710072, China)
|
| Abstract: |
| The existing research of process similarity mostly focuses on process structure but neglects the business semantic, and the similarity calculation on process structure is somehow deficient on computation complexity. To solve this problem, this paper proposes a business process representation with synthetical feature extraction, and a corresponding similarity calculation method is given at the same time. Weight notation of edges is used to extend the process structure for structure feature extraction based on the analysis of basic process control patterns, and high level business semantic is also involved for constructing business process semantic model for semantic feature extraction. The classic similarity metrics of node and edge are heuristically adapted to form a new structure similarity metric, and the similarity of business semantic is computed on vector space model and set theory. The total similarity is deduced by the weighted sum of structure and semantic similarity, and the computation model is self-adaptive to other existing methods by the adjustment of weight assignment. Finally, experiments are carried out to verify the performance of similarity computation, and the results show that the model of this paper is more adaptive and higher in computation efficiency when comparing to other methods in the literatures.
|
| Key words: business process model business semantic feature extraction process similarity process matching |
