Determinant Variables of Travel Time and Bus Travel Time Prediction: Case Study Ipoh-Lumut Corridor Intercity Regular Bus Service in Malaysia

This paper presents the determinant variables of travel time and bus travel time
prediction model. This study aims to assess some identified relevant variables and using
statistical model to develop adequate model for bus travel time prediction. Data were
collected by using GPS from regular bus system operated in rural with minimum congestion.
Two models were developed separately regarding to divided multiple-lane highway. The
results show that multiple linear regressions (MLR) were adequate to model bus travel time
prediction. Two models were significant based on analysis of variance (ANOVA) test at 95%
confidence limit. At case 1, MLR model for Ipoh to Lumut direction had mean absolute
percentage error (MAPE) of 14.8%. Meanwhile, at case 2, MLR model for Lumut to Ipoh
direction had MAPE of 12.1%. As the MAPE values of both models are small and reasonable,
therefore, models were adequate to be recommended for bus travel time prediction.