MACHINE LEARNING - BASED CASHLESS FARE SYSTEM USING RANDOM FOREST AND NEURAL NETWORKS FOR SMART URBAN MOBILITY
Abstract
Public transportation in Cebu City remains the primary mode of travel; however, manual fare collection creates inefficiencies such as inconsistent pricing, long waiting times, delayed boarding, and passenger-operator disputes. This study aimed to identify key stakeholder challenges and develop RidePay, an AI-enabled cashless fare collection system. A descriptive research design was used, involving 384 respondents, including commuters, drivers, conductors, and operators. Data were collected through structured surveys and analyzed using descriptive statistics. Findings revealed that 69.6% of commuters experience waiting times exceeding 30 minutes, while 68.8% prefer modernized transport systems. Additionally, all drivers and conductors rely on cash-based transactions, highlighting inefficiencies in fare management. The proposed system demonstrated high acceptability (M = 4.09), indicating strong stakeholder support. Developed using Agile methodology, RidePay integrates technologies such as Arduino, RFID, GPS, React, Node.js, and Firebase, with analytics supported by AI-driven insights. Results show that the system enhances efficiency, transparency, and operational performance, demonstrating its potential to support smart and sustainable urban transportation in Cebu City.
