LOGICAL - PROBABILISTIC REPRESENTATION OF CAUSAL DEPENDENCIES BETWEEN EVENTS IN BUSINESS PROCESS MANAGEMENT

Oksana Chala
2018 Advanced Information Systems  
The subject matter of the article are the processes of identifying knowledge in the form of causal relationships based on the analysis of the log of the business process. The goal is to develop a logical-probabilistic model of cause-effect relationships between pairs of log events that describes the implementation of the business process's action to support the solution of the task of automating the construction of the knowledge base of the process management system. Tasks: Select context
more » ... aints and limitations on the execution of business process actions that can be obtained as a result of log analysis; develop an approach to extract the probabilistic and logical components of cause-effect dependencies; to develop a logical-probabilistic model of causal relationships. The methods used are: methods for constructing predicate models; Bayesian methods of constructing probabilistic models. The following results are obtained. Formalized class of causal dependencies for knowledge-intensive business processes. Such dependencies can take into account informal knowledge of the business process. Within this class there are: a predicate description of the state of the context based on information about values of attributes of log events; contextual constraints on doing business process actions; probabilistic conditions for implementing the business process. Conclusions. The scientific novelty of the results obtained is as follows: a logical-probabilistic model of cause-effect relationships between pairs of log events describing the performance of the business process is proposed. The model binds a logical description of the state of the context before and after the completion of each activity of the business process, as well as a logical description of the constraints on the actions of the process and a probabilistic description of the conditions for the execution of these actions. In practical terms, the model provides an opportunity to solve problems of extracting, expanding and integrating knowledge based on the analysis of logs of business processes. K e ywor d s : causality; knowledge; knowledge base; dependencies; business process; workflow; event log.
doi:10.20998/2522-9052.2018.2.07 fatcat:mxpcicyhmjfyrbazo4qdtxhkcm