Peramalan Jumlah Penumpang Domestik di Tiga Bandara Utama Menggunakan -Generalize Space Time Autoregressive with Exonogeneous Variable (GSTARX)

Ahmada, Lailatul (2024) Peramalan Jumlah Penumpang Domestik di Tiga Bandara Utama Menggunakan -Generalize Space Time Autoregressive with Exonogeneous Variable (GSTARX). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003201054-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Transportasi udara adalah moda mobilisasi yang paling efektif dan efisien. Jumlah penumpang pesawat mengalami fluktuasi, dengan peningkatan tajam menjelang Idul Fitri, dan akhir tahun (Natal dan Tahun Baru), serta penurunan signifikan selama pandemi Covid-19. Oleh karena itu, peramalan jumlah penumpang sangat diperlukan untuk mendukung prediksi masa depan. Model ARIMA dan model GSTAR dikembangkan menjadi ARIMAX dan GSTARX dengan variabel eksogen seperti dummy tren, bulanan, variasi kalender, dan intervensi pandemi Covid-19. Studi kasus ini diterapkan pada tiga bandara utama di Indonesia: Soekarno-Hatta, Juanda, dan I Gusti Ngurah Rai. Penelitian ini bertujuan untuk mendapatkan model terbaik yang sesuai untuk peramalan jumlah penumpang di tiga bandara tersebut. Semua variabel dummy eksogen dimodelkan dengan regresi time series terlebih dahulu, model regresi ini digabungkan dengan model ARIMA sehingga didapat model ARIMAX, sementara untuk model GSTARX didapatkan dari memodelkan residual regresi time series. Hasil analisis menunjukkan bahwa model terbaik untuk Bandara Soekarno-Hatta dan I Gusti Ngurah Rai adalah model GSTARX dengan bobot invers jarak dikarenakan menghasilkan MAPE testing terkecil 6,085% dan 12,964%, sementara model terbaik untuk Bandara Juanda adalah model ARIMAX yang menghasilkan MAPE 18,322% yang merupakan 0,659% lebih rendah daripada model GSTARX dengan bobot invers jarak. Sehingga didapatkan kesimpulan bahwa model terbaik dalam meramalkan jumlah penumpang domestik keberangkatan pada tiga bandara utama adalah GSTARX dengan bobot invers jarak yang menghasilkan rata-rata MAPE sebesar 12,704 %. Prediksi selama 15 periode secara umum sesuai dengan pola data testing. Model GSTARX dengan bobot invers jarak memberikan efektivitas 46,026 % dibandingkan model ARIMAX. Ini menunjukkan bahwa dependensi spasial antara ketiga bandara utama berkontribusi besar pada jumlah penumpang domestik keberangkatan. Sementara peramalan dua periode mendatang (April &Mei 2024) yang menujukkan terjadinya kenaikan di ketiga bandara utama.
========================================================================================================================
Air transportation is the most effective and efficient mode of mobilization. The number of airplane passengers fluctuates, with sharp increases during Eid al-Fitr and the end of the year (Christmas and New Year), and significant declines during the Covid-19 pandemic. Therefore, forecasting the number of passengers is essential to support future predictions. The ARIMA and GSTAR models were developed into ARIMAX and GSTARX models with exogenous variables such as dummy unis, monthly variations, calendar variations, and Covid-19 pandemic interventions. This case study is applied to three major airports in Indonesia: Soekarno-Hatta, Juanda, and I Gusti Ngurah Rai. The research aims to find the best model suitable for forecasting the number of passengers at these airports. All exogenous dummy variables were first modeled with time series regression; this regression model was combined with the ARIMA model to obtain the ARIMAX model, while the GSTARX model was derived from modeling the residuals of the time series regression. The analysis shows that the best model for Soekarno-Hatta and I Gusti Ngurah Rai Airports is the GSTARX model with inverse distance weighting, as it produced the smallest testing MAPE of 6.085% and 12.964%, respectively. Meanwhile, the best model for Juanda Airport is the ARX model, with a MAPE of 18.322%, 0.659% lower than the GSTARX model with inverse distance weighting. Therefore, it can be concluded that the best model for forecasting the number of domestic departing passengers at these airports is the GSTARX model with inverse distance weighting, yielding an average MAPE of 12.704%. The predictions for the next 15 periods generally match the testing data pattern. The GSTARX model with inverse distance weighting provides 46.026% more effectiveness compared to the ARX model. This indicates that the spatial dependency between the three major airports significantly contributes to the number of departing domestic passengers. Meanwhile, the forecasts for the next two periods (April & May 2024) show an increase at the three major airports.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	GSTARX, number of domestic air passengers, spatio-temporal, calendar variation, intervention, ARIMAX, time series, jumlah penumpang pesawat domestik, spatio temporal, variasi kalender, intervensi, ARIMAX, deret waktu
Subjects:	Q Science > QA Mathematics > QA274.2 Stochastic analysis Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA280 Box-Jenkins forecasting
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Physics > 45201-(S1) Undergraduate Thesis
Depositing User:	lailatul ahmada
Date Deposited:	08 Aug 2024 12:28
Last Modified:	08 Aug 2024 12:28
URI:	http://repository.its.ac.id/id/eprint/114745

Actions (login required)

View Item