基于图挖掘的服务流程推荐方法研究

The demand of dynamics, uncertainty, individuation and inter-organization for business process modeling is more and more exigent in modern enterprises recently, this project tries to explore the theory and algorithm of business process recommendation based on graph mining, applying business intelligence and service computing technologies to process modeling. In the project ,we will focus on pattern extraction and distance computation, and mainly studies the frequent process pattern mining algorithm, distance between processes, business activities recommendation algorithm and other key technologies. Then, we discover associate rules between business activities and processes, and feedback these rules to process designers with a personalized recommendation service, and builds a theoretical system of workflow recommendation to improve the effectiveness, accuracy, intelligence and generality of process modeling. In addition, the project will verify the models and algorithms through developing one recommendation platform and applying the research results into three typical application scenarios including process modeling, service composition and inter-organizational service.

课题针对现代企业业务流程建模存在的动态性、个性化、不确定性、跨界性等迫切需求，将商业智能、服务计算技术应用到业务流程建模中，研究基于图挖掘的服务流程推荐理论问题和计算方法。课题围绕"模式提取"与"距离计算"两大核心问题展开研究，重点研究业务流程数据集中的频繁模式挖掘算法、流程间距离分析模型以及业务活动推荐算法等关键技术，发现业务活动或服务流程在相似应用环境中的关联规则等隐含知识，并将这些知识反馈给建模人员，为建模人员推荐个性化服务流程，形成流程推荐技术的理论体系，大幅度提高流程建模的高效性、准确性、智能性和普适性。通过研制一个基于图挖掘的服务流程推荐平台,在流程建模、服务组合和跨界服务三个典型应用场景下进行仿真实验与真实系统应用，对相关模型与算法进行验证。

项目针对现代企业业务流程建模存在的动态性、个性化、不确定性、跨界性等需求，将商业智能、服务计算技术应用到业务流程建模中，研究基于图挖掘的服务流程推荐理论问题和计算方法。项目围绕“模式提取”与“距离计算”两大核心问题展开，重点研究业务流程数据集中的频繁模式挖掘算法、流程间距离分析模型以及业务活动推荐算法等关键技术，形成了流程推荐技术的理论体系，大幅度提高了流程建模的高效性、准确性、智能性和普适性。最后，项目研制了一个基于图挖掘的服务流程推荐平台，在流程建模、服务组合等典型应用场景进行仿真实验与真实系统应用，并对相关模型与算法进行了验证。