Incident prediction is critical in industrial settings to prevent disruptions and optimize operations by anticipating failures. To this end, equipment logs are commonly utilized and converted into an XES-like log format, wherein each event is associated with a single case object (i.e., equipment) reporting its status. However, the resulting event log overlooks other data sources, such as logs of pre- and post-incident processes. These logs report activities applied to the equipment's related objects (e.g., hardware, software) in non-structred way, presenting challenges when using the XES-like log format. This paper introduces a meta-model to integrate and represent these diverse data sources in formalized format. To this end, we propose a domain-specific object log tailored to the incident-monitoring domain to represent multiple equipment-related objects. This meta-model was validated with a real-life dataset, showing how it provides an effective and structured representation of incident-related data.

Summary At the stage of identifying and modeling business processes (BP), traditional methods for collecting business expertise are time consuming since they mainly rely on physical communication. Moreover, there are some BP with no formal documentation and their execution relies on implicit knowledge of workers. Therefore, analyzing the logs data generated by messaging systems allowing workers to share such implicit knowledge has been a research area of growing interest. BP knowledge introduced by workers in emails could actually cover multiple BP perspectives that would vary from activity, actors, and data perspectives. This article gives an overview of the current state‐of‐the‐art in multi‐perspective BP discovery from messaging systems as well as the main existing research questions. Different approaches are presented, discussed, and compared against each other on a qualitative level deduced from the defined set of research questions. The resulting overview covers both advantages and disadvantages of current techniques and the research questions that have not been addressed to date. This should provide a solid basis for further research.

International audience

At the stage of identifying and modelling business processes (BP), traditional methods for collecting business expertise are time consuming since they mainly rely on physical communication. Moreover, there are some BP with no formal documentation and their execution relies on implicit knowledge of workers. Therefore, analyzing the log data generated by messaging systems allowing workers to share such implicit knowledge has been a research area of growing interest. This paper gives an overview of the current state-of-the-art in BP and activity discovery from messaging systems as well as the main existing research questions. Different approaches are presented, discussed and compared against each other on a theoretical level deduced from the defined set of research questions. The resulting overview covers both advantages and disadvantages of current techniques and the research questions that have not been addressed to date. This should provide a solid basis for further research.

In pervasive environments, end-users have heterogeneous devices to access their different services. These services are usually distributed over different devices and each service should be able to run in the most appropriate device. However, current technologies do not address the integration between these services and as a consequence the end user does not have the possibility to access the services in an optimal way. In this paper, we define and implement new mechanisms that enable a seamless integration of a service in the end-user pervasive environment. First, these mechanisms enable the end-user to personalize his/her pervasive environment by running each functionality in the most suited (preferred) device, and then to make these services communicate with each other. The specificity of our approach is that we split each application into independent functionalities, and then, we define and implement on the end-user devices a distributed mechanism that detects automatically semantic compatibilities between these functionalities.

Widget aggregators such as iGoogle and Netvibes are broadly adopted by the mass market. They enable end-users to personalize their environment with their preferred services (Widgets). However, the usage in an enterprise context is not yet investigated. In this paper, we firstly show that in addition to personalization capability, the integration of business processes should be considered. Secondly, we propose a new Widget aggregator that enables the end-user to personalize a business process by chaining Widgets according to his/her needs and habits. Thirdly, we introduce a new approach for specifying an end-user process; an approach which enables even ordinary end-users, without computing skills, to define their processes. Finally, we validate these concepts by implementing and testing a prototype. As a consequence, this work does not only impact Widget aggregators, but it also innovates in end-user service creation research by proposing an intuitive tool, understandable even by ordinary end-users, for specifying their processes (composite services).

In this paper, we investigate the various service composition mechanisms and provide the impact of each of them on user-centric service development issues. We classify service composition mechanisms into three categories: automatic service composition, semi-automatic service composition, and static service composition. As services are today mainly driven by the user's needs, the following survey essentially focus on automatic service composition and semi-automatic service composition. This enables users to conceive their own personalized applications.