A simple process model generation using a new association rule mining algorithm and clustering approach

Abstract

In the recent research area, the association rule mining is one of the popular technique in the domain of data mining. The association rule mining is first implemented in dyeing unit in my previous research papers. But already this association rule mining technique is used in the area of finance, healthcare, automobile and sales and distribution, etc. In this article, the dyeing process, generate the simple process model for the dyeing unit. The simplified process model is not in the form of diagram, instead rules. These rules are developing using association rule mining algorithms. The process mining algorithms, Heuristic Miner (HM) and Disjunctive Workflow Schema (DWS) are used to generate the association rule mining rules. Hence, the proposed LinkRuleMiner (LRM) association rule mining algorithm is implemented in the dyeing unit using HM or DWS algorithm. The dyeing process is dynamic and unstructured in nature. The dyeing process is recorded and stored in the form of event logs. These event logs are converted in to the log file. These log files are given as input to the LRM algorithm. The LRM algorithm produces the simple association rules. These rules can be easily understood by the dyeing expert called dyer to process the colouring process of the dyeing unit. These generated association rules can also be grouped for the same or different shades using the clustering approach. Therefore, this article simplify the dyeing process using LRM algorithm to give better knowledge to the dyer and to reduce the dyeing processing problems of the dyeing unit.