流程监控与评估中多元数据整合研究

Big data is changing our's life and business environment, especially it brings organizations great business value,however, most of organizations has difficulty in mining knowledge and using big data.Manager and Knowledge workers usually lose themselves in the big, complex,low correlated and multi-sources data, thus it is important for them to aggregate multi-sources data before mining knowledge from big data. Research for big data attracts great volume of attention from both scholars and business organizations, however, few research has focus on multi-sources data aggregation. Therefore, in this project,in the context of business process monitoring and evaluation,how to choose data among multi-sources data for a particuliar business problem and position for organizing those chosen data is studied from domain knowledge and theory perspective, and then on the basis of which, techiniques for these chosen data aggregation is developed including Semantic ontology, knowledge bases, and algorithms for data analysis. Finally, simulation and case study are conducted for evaluating the whole methodology proposed in the project for multi-sources data aggregation. The expected contributions are as follows: firstly, with the domain knowledge and theory,the difficulty in multi-sources data aggregation caused by complex characteristic including low-correlated,heterogenous is alleviated by developing Semantic ontology,knowledge bases, related data analysis algorithms;secondly, business process monitoring and evaluation is supported by aggregating business process logs,context and Internet user generated content,based on which, business process monitoring and evaluation can enahnce organization's capability in serving customers and make organizations more adaptable to dynamic enviroment;Thirdly, the methodology provided in this project benefits big data utilization and promotes knowledge quality mined from big data.

本项目立足于企业缺乏大数据价值有效挖掘的现状,多元数据整合为大数据有效挖掘前提的事实,针对多元数据复杂性即互相隔绝,类型异构,关联度较低造成了数据整合难的问题以及目前学术界较为缺乏多元数据整合研究内容的现实,提出了以流程监控与评估为研究背景,首先利用领域知识与相关理论分析数据整合时对多元化数据内容的选择,以及数据整合位置点的选择,然后开发设计数据整合所需的语义知识模型,知识库,数据模型,数据分析算法等技术,并应用仿真与实际数据案例对整合方法进行评估.项目为大数据有效挖掘奠定了理论基础,同时还开发了相关支持技术,为流程监控与评估和其他商业领域提供了有效利用大数据的理论与技术方案.

本项目以流程监控与评估为研究背景，利用领域知识与相关理论对大数据中多元化数据整合位置点进行了选择分析。利用本体与知识推理等技术，建立了上下文数据与流程日志这些异构的多渠道数据的整合模型。在整合后的数据上开发了相应的数据分析算法。数据分析的最终结果揭示了流程活动执行人的工作动机，合作模式等特点以及这些特点对流程效率的影响。研究成果不仅为流程提供了管理建议，同时为大数据多元整合提供了理论与技术支持。项目组成员在执行项目期间，因以数据的角度来看流程的监控与评估受到启发，对项目的研究内容做了扩展，为传统的流程设计模型增加了数据存储功能，使得流程设计模型以数据为中心能够被重复利用。同时由于供应链质量监控与流程监控与评估有着相似性，项目组成员以数据为中心的角度为供应链质量监控建立了数据模型，该模型能够帮助管理人员动态监控委托代理合同在执行期的失效问题。在项目实施过程中， 发表学术论文5篇，其中3篇SCI/SSCI检索，2篇EI检索，1篇发表在《Decision Support Systems》，1篇发表在《Electronic Commerce Research》，1篇发表在《Journal of Information Science》, 1次在会议PACIS 2016做了海报张贴；培养了3名相关方向硕士研究生。