APPLICATION OF GENERALIZED SPACE-TIME AUTOREGRESSIVE MODEL ON THE NUMBER OF VEHICLES THAT EXIT THE PURBALEUNYI TOLL GATES WITH TIME CORRELATED ERRORS

• Main Author: TAZKI NABILAH (NIM: 10114044), FAKHITA
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/27124
• Institution: Institut Teknologi Bandung

Toll Gate is a vital place for the entry of vehicles from within and outside the city. At any time there are differences in the number of vehicles entering and exiting from each toll gate. This shows the influence of time and location so it can be analyzed with GSTAR (Generalized Space-Time Autoregressive) model. In general, the error in the GSTAR model is assumed to be normally distributed and independent. However, in reality these assumptions are difficult to fulfill so that they are weakened with time correlated errors. In this final project used secondary data that is data of the number of vehicles which exit from eight Purbaleunyi toll gates. Later, on the data applied the GSTAR model (1;1) assuming independent errors and time correlated errors following the martingale difference process. After comparison between the original data and the estimation results, it was found that the error followed the martingale difference process. Therefore, it can be concluded that the GSTAR (1;1) model with time correlated error is a suitable model for the data of many vehicles out of eight toll gates.