Comparing the Performance Rating Between Neutral Network and Autoregressive Integrated Moving Average Model for Vehicle Traffic Prediction Estimation
Abstract
Traffic congestion poses significant challenges worldwide, resulting in lost hours
of travel time and increased fuel consumption. Accurate traffic prediction is
crucial for mitigating these issues. Traditional traffic prediction approaches such
as ARIMA are limited by their inability to handle large datasets, inaccurate
predictions, and time constraints. This study explores the use of Feed Forward
Back Propagation Artificial Neural Network (FFBPANN) in the determination of
traffic prediction. Five different FFBPANN architectures were created to
determine the optimal topology. The results show that the architecture with 20
hidden layer neurons achieved the best performance, with a mean square error of
0.19188 at 5 epochs. The implementation of FFBPANN can enhance traffic
management systems, reduce congestion, and improve urban mobility, ultimately
contributing to more efficient transportation networks.