Peak-Hour Urban Road Traffic Volume Prediction Model Based on Short-Term Traffic Count in Putrajaya

The number of vehicles has risen rapidly as transportation demand increases. These challenges have led to growing concerns over traffic volume due to the escalation of population and urbanisation. Traffic count data is vital for planning and management, understanding present and future transportation demands, guiding infrastructure enhancements, and assessing existing transportation systems. However, data collection and analysis take time, which is a drawback of traffic count. Therefore, this study focuses on developing a peak-hour traffic volume prediction model based on short-term traffic counts. This study was conducted at seven arterial roads in Putrajaya. The traffic counts survey was conducted for 3 hours during peak times in the morning and evening. Besides, the vehicles were divided according to their classes during the traffic count process. Statistical regression analysis, including Single Linear Regression and Multiple Linear Regression, was used to develop the prediction models. A series of analyses were conducted to create an optimal model for predicting traffic volume during peak hours. After developing the model, validation against manual traffic counts was performed using Mean Absolute Percentage Error (MAPE) to assess the model's accuracy. If the percentage error is low, the model can anticipate peak hour volumes based on short-term data. In conclusion, the analysis of the result shows that the peak hour road traffic volume prediction model based on a 15-minute short-term traffic count using MLR analysis is the best model to use to predict traffic volume in Putrajaya. Furthermore, the findings indicate that the model has 12.09% MAPE, which suggests it closely follows actual traffic volume data.