ANALYSIS OF NEURAL NETWORK AND TIME SERIES MODELS FOR PREDICTION OF THE NUMBER OF VEHICLES ENTERING BANDUNG CITY THROUGH THE PASTEUR TOLL

• Main Author: Rakaditya, Difa
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39141
• Institution: Institut Teknologi Bandung

Prediction of traffic information is a complicated problem because of the many factors that influence it. Toll roads are often the main focus in predicting traffic information. Data on the number of vehicles on toll roads is arranged based on the time of observation, thus forming a time series data. To find out information on traffic in the future, it takes information from the past to describe the time series. The mathematical model used to predict time series data is ARIMA and Seasonal ARIMA. Along with technological developments, alternative methods emerged such as artificial neural networks and recurrent neural networks. In this final project a comparative analysis is carried out between Seasonal ARIMA models, artificial neural networks, and recurrent neural networks on data on the number of vehicles entering Bandung via Pasteur Toll Road. The neural network model produces more satisfying results based on mean absolute error value produced.