Data-Driven Analysis of Loan Approval Service of a Bank using Process Mining

This research presents the application of Process Mining in analyzing the loan approval service process of a branch of a bank in Thailand using the Fuzzy Miner and Social Network Miner algorithms through a data-driven discovery of process bottlenecks to improve the efficiency of the banking service. Three steps are required to proceed: 1. Specifying the type of the event log dataset. 2. Importing data into the process mining tool Fluxicon Disco, which runs based on the Fuzzy Miner algorithm. 3. Importing data into the Rapid Miner platform: Three results were modeled to be obtained as follows: 1. Statistical analysis of the results, 2. Fuzzy models/graphs (in terms of Frequency and Time Performance metrics), and 3. Social Network graphs/models (regarding Handover of Work metric). According to the model, the analysis process was repeatable for 93.4% of the analysis process and used three workers for 60.80% of the system while taking an average of 11 hours. Based on the real-event log, we suggested that the bank admins evaluate their control over the distribution of the staff workload to consider hiring more individuals for the position of load approval analysis. In general, the approach proposed in the paper develops a method for applying for loan approval quickly in a timelier manner.