Research On The Method Of Mining Configuration Information Based On The Behavioral Profiles Of Petri Net

Business process mining which including the design and implementation of process, building and anglicizing of the model and the execution of program is a new type of application in business process management. Business process mining through analysis event logs which recorded by information system to establish process model according to the running of business process, then the original model is optimized by using the log sequences and obtain the final model. The research of process mining technology not only helps to reduce operating costs and improve the efficiency of the system, but also plays an important role in improving the production efficiency and improving the quality of service.As business process management becomes more and more important, business process requirements are also relatively improved a lot. The existing process mining techniques mainly for mining workflow model, but in the process of mining business process model will often appear some transitions that exist in the process of the model but not appear in the log sequence. What is more, there is a lot of the model that include hidden transitions in the reality. Therefore mining the hidden transitions from the event logs can be build process model more accurate and increase efficiency of the operation that is conducive to improving production and service efficiency. The current method for mining process model configuration information such as hidden transitions is based on algorithm a. However there are some disadvantages, such as large amount of computation and the accuracy of hidden transitions’ mining. A new method of mining hidden transitions is proposed in this article that is based on behavioural profiles and probabilistic behavioral relations to analysis the log sequences, then catch the process model with configuration information which make model more complete. The main contributions of this paper are followed.(1) A Mining Method of the Hide Transitions from the Event Logs Based on the Behavioral Profiles of Petri Net. It is a difficulty that mining hidden transitions from the event log in the research of the process model. There are some researches of mining hidden transitions that used in the free choice nets, but it is not suitable for the complex process model. Then the method of mining hidden transitions based on behavioural profiles is introduced as fellow:Firstly, select the event logs which had the highest frequency to catch the behavioral profiles and establish the initial model. Second, regard the event logs with highest frequency in remainder log as increment to optimize the initial model. Repeat the above steps until the hidden transitions are found through the differences of behavior profile. Finally, the fitness and appropriateness of the process model with configuration information are calculated and used to test the model.(2) Mining the Model with the Hidden Transitions Based on Petri Net Probabilistic Behavioral Relations. In the mining process, it is one of the main challenges in the process mining that to find a possible best process model based on the observed behavioral. The hidden transitions refer to some transitions which existing in the process of the model, but no changes in the log sequence. Considering the influence of hidden transitions on the mining process model, the method of mining process models with configuration information based on probabilistic behavior relation form incomplete event logs is proposed in this paper. Firstly, according to the probabilistic behavior relation obtain the relationship between modules, and then calculate the probability that the relationship between the transitions. The process model with configuration information is established based on the relationship. Finally, the specific locations of the hidden transitions are determined by analyzing the default transitions in the log sequence. The method can be effectively and quickly to mine the positions hidden transitions. The experiment proved that the accuracy of this method.