Object-centric event data is a generalization of case-centric event logs that allows events to be associated with multiple objects of different types. This flexibility allows extracting, modeling, and analyzing real-life processes more accurately by including multiple perspectives. Publicly available datasets are needed to support the development, evaluation, and conceptualization of new object-centric process mining concepts and techniques. In this paper, we present a generic framework to extract public scientific publication data into object-centric event logs. Apart from the publications and their authors, our framework also integrates information on the topics and keywords of works, as well as scientific conferences in the context of which research papers were published or presented. The extraction approach is generic and can extract available works for any given author or keyword, yielding a single object-centric event log as a result. The resulting dataset can then be used in object-centric process mining techniques, as well as for other general analysis, as it contains information on a plurality of different aspects. To demonstrate the feasibility of our approach, as well as the analytical potential of the resulting logs, we present a case study using the publication works of Wil van der Aalst. For example, we showcase how models mined by object-centric process model discovery describe the control-flow of the obtained data and analyze the popularity and development of keywords. The case study’s analysis artifacts and the extracted object-centric event log, as well as the extraction framework code, are all publicly available.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Object-Centric Extraction and Analysis of Scientific Publications

  • Aaron Küsters,
  • Cameron Pitsch,
  • Christian Rennert,
  • Jan Niklas van Detten,
  • Ali Norouzifar,
  • Viki Peeva,
  • Tian Li,
  • Benedikt Knopp

摘要

Object-centric event data is a generalization of case-centric event logs that allows events to be associated with multiple objects of different types. This flexibility allows extracting, modeling, and analyzing real-life processes more accurately by including multiple perspectives. Publicly available datasets are needed to support the development, evaluation, and conceptualization of new object-centric process mining concepts and techniques. In this paper, we present a generic framework to extract public scientific publication data into object-centric event logs. Apart from the publications and their authors, our framework also integrates information on the topics and keywords of works, as well as scientific conferences in the context of which research papers were published or presented. The extraction approach is generic and can extract available works for any given author or keyword, yielding a single object-centric event log as a result. The resulting dataset can then be used in object-centric process mining techniques, as well as for other general analysis, as it contains information on a plurality of different aspects. To demonstrate the feasibility of our approach, as well as the analytical potential of the resulting logs, we present a case study using the publication works of Wil van der Aalst. For example, we showcase how models mined by object-centric process model discovery describe the control-flow of the obtained data and analyze the popularity and development of keywords. The case study’s analysis artifacts and the extracted object-centric event log, as well as the extraction framework code, are all publicly available.