Applying Fuzzy-Genetic mining in conformance and dependency relations

In this paper, we used genetic algorithm to discover a Petri net given an event log received from one of the universities in Thailand. The event log contained information about 299 cases and 569 activities. For each case, the performed tasks - in regard to students' registration process - and the moment of completion were recorded. In short, using Genetic algorithms made us capable of simulating the control flow perspective of students' registration process in the case study. In contrary with Alpha-algorithm and Heuristic Miner approaches, the problem of noise is naturally tackled by the genetic algorithm because, per definition, these algorithms are robust to noise. Yet the main challenge in a genetic approach is the definition of a good fitness measure because it guides the global search performed by the genetic algorithm. Therefore, initially, this paper explains how the genetic algorithm works and then, experiments with real-life registration log shows that the fitness measure indeed leads to the mining of process models that are complete (can reproduce all the behavior in the log) and precise (do not allow for extra behavior that cannot be derived from the event log). Consequently, the discovered graphical models were depicted and compared in Genetic and Fuzzy environments.