A meta-model for operational support

Process mining techniques can be used to extract information from event logs helping organizations to gain insights into the operation of their business processes. It is now possible to provide such analysis information for ongoing executions in an on-line setting, yielding predictions about the current execution or recommendations of actions to take, commonly referred to as operational support. This paper defines a meta-model for operational support making it possible for users or tools to receive operational support without detailed knowledge of the algorithm providing answers. The meta-model defines four types of queries: A simple query asks diagnostic information about the current execution and a compare query compares the current execution with other similar executions. A predict query considers the future of previous executions similar to the current one to make a prediction about its future. A recommend query recommends the best next action to take based on predictions. We formally define our meta-model and take care to provide implementation suggestions for each defined concept. We present our implementation of the meta-model in the work flow system Declare and the process mining framework ProM.