Research On Business Process Discovery From Relational Database Operations

The execution of business processes generates a large number of logs,which provides an important basis for business process discovery.Process discovery is one of the three major research directions of process mining,which aims to dig out business model that can describe process behavior from logs accurately.The process finds that a log in a standardized format is required,but in many cases,the information system does not provide a log in a usable form,which requires a combination of data storage mechanisms to obtain logs.Existing log extraction and conversion methods lack support of key business data.This article constructs logs by extracting relational database data operation statements,and then discovers the business process model.The main contents of this paper are as follows.Firstly,in order to obtain eventlogs that can be used for process discovery,a SQLEvent Log conversion algorithm is proposed.The SQL statement sequence obtained from information system is reorganized and sorted by timestamp to obtain SQL mapping sequence.Analyze the structure of the SQL mapping sequence,extract the M-R-F-T attributes of each sentence,and construct the SQLEvent Log of the business process.Secondly,combining the characteristics of control flow and data flow in M-R-F-T attributes,a business process discovery algorithm based on SQLEvent Log is proposed.Define the SQLEvent attribute tag value,encode and de-duplicate SQL instance sequence.Given the modeling threshold,construct SQLTrace sequence,combine relational database operation with Petri net,and establish SQL-net model for business process.Finally,the above algorithms are verified through simulation data experiments.Calculate the average consumption time of the business process trace the fitness of the business process model’s adjacent events and the trace fitness,evaluate the business process found,and verify the rationality of the business process model and the effectiveness of the algorithm.