Workflow log repeated task recognition method based on relation matrix

Abstract

The invention discloses a workflow log repeated task recognition method based on a relation matrix. According to the invention, a strategy that log events are firstly classified and then clustered isadopted, in a division process, by finding a circulation and parallel structure with a stable structure, division types are reduced; in a clustering process, according to event contexts, similarity ofevery two repeated tasks in a candidate repeated task set is calculated; in the clustering process every time, model quality is calculated; and finally, after the clustering is finished, the optimalclustering scheme is selected, wherein each cluster represents one real different task. According to the invention, accuracy of following workflow model excavation can be improved.

Description

technical field [0001] The invention belongs to the technical field of data mining, and in particular relates to a method for identifying repetitive tasks in a workflow log based on a relationship matrix. Background technique [0002] With the wide application of intelligent manufacturing and business process management, workflow technology has developed rapidly in recent decades, and is constantly evolving towards intelligence and flexibility. Workflow technology abstracts the process in daily life into a process model, decomposes work into tasks and assigns them to different roles, executes and supervises according to the established process, and its application purpose is to improve work efficiency. Workflow model mining is to extract the process model by analyzing the operation logs in the information system. Compared with the artificially designed process model, it is more objective and can bring great significance to the enterprise to discover, monitor and improve the ...

AI Technical Summary

Problems solved by technology

[0003] During the process modeling process, the modeler will assign the same activity name to different task nodes to represent the execution of the same activity at different stages of the process, and the event name recorded in the running log is usually determined by the activity name rather than the task Name, which makes it impossible to determine whether events with the same name (duplicate events) appearing in the track represent the same task or a different task in the case of duplicate tasks

E.g figure 1 As shown in (a), the original model of a hospital’s diagnosis and treatment process includes multiple diagnostic operations c, blood test operation b, and X-ray detection operation x. If the model is directly mined based on these original logs, for example figure 1 (b), it may be possible to dig out non-existent and meaningless trajectories in the original diagnosis and treatment process due to the repetition of activity names, and it is impossible to restore the real diagnosis and treatment process

[0004] However, the premise of most existing workflow model mining algorithms is that there is a one-to-one correspondence between the events recorded in the running log and the tasks in the model.

When these algorithms are used for process mining of running logs containing repetitive tasks, the obtained process models often have too many connected nodes and meaningless loop structures, resulting in reduced accuracy of model restoration

Images