PERBANDINGAN MAXIMUM LIKELIHOOD ESTIMATION (MLE) DAN BAYESIAN INTEGRATED NESTED LAPLACE APPROXIMATION (INLA) PADA PERAMALAN TIME SERIES

Dewi Nur Sinta Lestari, NIM.: 18106010009 (2022) PERBANDINGAN MAXIMUM LIKELIHOOD ESTIMATION (MLE) DAN BAYESIAN INTEGRATED NESTED LAPLACE APPROXIMATION (INLA) PADA PERAMALAN TIME SERIES. Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (PERBANDINGAN MAXIMUM LIKELIHOOD ESTIMATION (MLE) DAN BAYESIAN INTEGRATED NESTED LAPLACE APPROXIMATION (INLA) PADA PERAMALAN TIME SERIES) 18106010009_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf - Published Version Download (3MB) | Preview
	Text (PERBANDINGAN MAXIMUM LIKELIHOOD ESTIMATION (MLE) DAN BAYESIAN INTEGRATED NESTED LAPLACE APPROXIMATION (INLA) PADA PERAMALAN TIME SERIES) 18106010009_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf Restricted to Registered users only Download (4MB) | Request a copy

Abstract

Forecasting is one method for estimating future aircraft passenger. The time series method, combined with the MLE approach and the Bayesian INLA approach, provides an alternative solution for data modeling. The purpose of this rese is to compare the MLE approach and the Bayesian INLA approach to time series forecasting using data from Soekarno-Hatta Airport. The data used ranges from January 2018 to December 2021. The best model criteria in the time series method using the MLE approach are Akaike's Information Criterion (AIC), whereas the Bayesian INLA approach uses Deviance Information Criterion (DIC) and Wanatabe-Akaike Information Criterion (WAIC). The ARMA(2,2) model is the best model for data on the number of airplane passengers at Soekarno-Hatta Airport using the MLE approach, while the AR(1) model with penalized complexity (PC) prior is the best model for data on the number of airplane passengers at Soekarno-Hatta Airport using the INLA approach. For the case of data on the number of aircraft passengers at Soekarno-Hatta Airport, the time series method with the INLA approach has a lower MAPE value than the time series method with the MLE approach based on the MAPE value from the prediction results of the two methods. The MAPE value of the INLA approach is less than 10 percent, indicating that it produces very accurate forecasting results.

Share this knowledge with your friends :

Actions (login required)

View Item