Smart Ticketing: Application of Random Forest for Bus Ticket Price Prediction

Tickets are essential in public transportation because they serve as proof of payment that the ticket holder has the right to participate in the bus trip. This research built a bus ticket sales system by applying the Random Forest algorithm. We use this algorithm to classify and predict bus ticket prices based on bus facility parameters, distance, and fuel prices. This system has a prediction feature to help bus managers set suitable prices based on these parameters. Implementation of the bus ticket purchasing and selling system can increase the efficiency of managing the sale and purchase of bus tickets, such as buyers being able to check and order the departure schedule directly, choose a seat on the bus without having to come right away to the sales counter, and choose the appropriate bus variety with the buyer's intentions. We collected this research data from ORGANDA, a bus management organization, and several bus ticket counters. We obtained data for the last five years. We tested the random forest algorithm by comparing the actual price with the system's predicted price, which was derived from the algorithm. The average prediction accuracy obtained from several tests reached 89.52%. The system implementation is based on a website using the Laravel and Django frameworks.