Fraud detection on event logs of goods and services procurement business process using Heuristics Miner algorithm

摘要

Event logs are history records that contain sequence data for the activity of a case that has been executed by an information system. Event logs can be valuable information with a technique called mining process. With this technique, cheating on the business processes of an enterprise can be detected early on. Thus, the company can commit further examination of business processes, especially the business process of procurement of goods and services to achieve business process is expected.[8] In this study, management data of event log obtained from log data at each event transaction procurement and services. The event log data is then analyzed using a heuristic miner algorithm. Heuristics miner algorithm chosen because it has advantages that are not owned by Alpha++ algorithm that this algorithm can calculate the frequency relation between activities in the log to determine the causal dependency. Heuristic Miner can be used to determine the predominant process of thousands of logs and detect behaviors that are not common in a process.[11] This study aims to detect anomalies on business processes that occur during the process of procurement of goods and services by calculating the fitness value of the event log into the system. Heuristic miner algorithm using the results obtained identification accuracy of 0.88%.