Fraud detection of bus ticket sales by using spatio temporal data mining

Abstract

Cheating in the sale of bus tickets is often found in transportation service providers who still use the conductor as a ticket seller on the bus. The high cost of supervision, lack of honesty, unification of the sales function and the ticket control function on the conductor makes this fraudulent practice a problem that companies must handle. By looking at the behavior of ticket sales for each individual through the method of spatio temporal clustering can detect fraudulent behavior that occurs. The bus ticket sales deception process is implemented in Bogor's Bus Rapid Transit (BRT). The results show that there are 3.2% of high-potential officers cheating ticket sales. By knowing the cheating behavior of ticket sales, the company can follow up with the policy so that cheating behavior does not become the company culture.