Klasifikasi Keterlambatan Keberangkatan Pesawat Domestik dengan Metode LSTM dan XGBoost (Studi Kasus: Bandara Sultan Aji Muhammad Sulaiman Balikpapan) - Submit Jurnal

Aryanti, Sahda (2025) Klasifikasi Keterlambatan Keberangkatan Pesawat Domestik dengan Metode LSTM dan XGBoost (Studi Kasus: Bandara Sultan Aji Muhammad Sulaiman Balikpapan) - Submit Jurnal. Bachelor thesis, Institut Teknologi Kalimantan.

[img] Text
11211080_cover.pdf
Restricted to Registered users only until 4 October 2027.

Download (139kB) | Request a copy
[img] Text
11211080_statement_of_authenticity.pdf
Restricted to Registered users only until 4 October 2027.

Download (369kB) | Request a copy
[img] Text
11211080_publishing_agreement.pdf
Restricted to Registered users only until 4 October 2027.

Download (396kB) | Request a copy
[img] Text
11211080_approval_sheet.pdf
Restricted to Registered users only until 4 October 2027.

Download (330kB) | Request a copy
[img] Text
11211080_preface.pdf
Restricted to Registered users only until 4 October 2027.

Download (262kB) | Request a copy
[img] Text
11211080_abstract_id.pdf
Restricted to Registered users only until 4 October 2027.

Download (174kB) | Request a copy
[img] Text
11211080_abstract_en.pdf
Restricted to Repository staff only until 4 October 2027.

Download (159kB) | Request a copy
[img] Text
11211080_table_of_content.pdf
Restricted to Repository staff only until 4 October 2027.

Download (326kB) | Request a copy
[img] Text
11211080_illustrations.pdf
Restricted to Repository staff only until 4 October 2027.

Download (247kB) | Request a copy
[img] Text
11211080_tables.pdf
Restricted to Repository staff only until 4 October 2027.

Download (239kB) | Request a copy
[img] Text
11211080_chapter_1.pdf
Restricted to Repository staff only until 4 October 2027.

Download (442kB) | Request a copy
[img] Text
11211080_chapter_2.pdf
Restricted to Repository staff only until 4 October 2027.

Download (969kB) | Request a copy
[img] Text
11211080_chapter_3.pdf
Restricted to Repository staff only until 4 October 2027.

Download (733kB) | Request a copy
[img] Text
11211080_chapter_4.pdf
Restricted to Repository staff only until 4 October 2027.

Download (2MB) | Request a copy
[img] Text
11211080_conclusions.pdf
Restricted to Repository staff only until 4 October 2027.

Download (179kB) | Request a copy
[img] Text
11211080_bibliography.pdf
Restricted to Registered users only until 4 October 2027.

Download (359kB) | Request a copy
[img] Text
11211080_enclosure.pdf
Restricted to Repository staff only until 4 October 2027.

Download (482kB) | Request a copy
[img] Text
11211080_paper.pdf
Restricted to Repository staff only until 4 October 2027.

Download (1MB) | Request a copy
[img] Text
11211080_presentation.pdf
Restricted to Repository staff only until 4 October 2027.

Download (3MB) | Request a copy
[img] Text
11211080_Form. TA-020.pdf
Restricted to Repository staff only until 4 October 2027.

Download (390kB) | Request a copy

Abstract

Sebagai negara kepulauan, Indonesia menjadikan transportasi udara sebagai pilihan utama untuk perjalanan jarak jauh. Seiring meningkatnya permintaan, industri penerbangan mengalami berbagai permasalahan operasional, salah satunya keterlambatan penerbangan. Bandara Sultan Aji Muhammad Sulaiman (SAMS) Balikpapan juga mengalami permasalahan keterlambatan penerbangan, dan belum memiliki sistem prediksi keterlambatan yang akurat. Penelitian ini bertujuan membangun model klasifikasi untuk memprediksi keterlambatan keberangkatan pesawat di Bandara SAMS, dengan mengombinasikan metode Long Short Term Memory (LSTM) dan Extreme Gradient Boosting (XGBoost). Dataset yang digunakan penelitian ini sebanyak 9.855 riwayat penerbangan maskapai dengan periode 10 Juli 2024 - 10 Januari 2025 dan menggunakan evaluasi matriks accuracy, precision, recall, F1-score, serta AUC. Model menggunakan teknik Random Over-Sampling untuk menyeimbangkan data training serta windowing untuk mengolah fitur temporal. Pemisahan data model menggunakan startified split dan balanced split (Data Testing). Fitur temporal hasil ekstraksi LSTM kemudian digabung dengan fitur non-temporal sebagai input XGBoost. Hasil menunjukkan bahwa model kombinasi LSTM dan XGBoost dengan balanced split pada rasio 90%:10% menghasilkan kinerja terbaik, dengan precision menapai 0.92189, recall sebesar 0.91899, f1-score bernilai 0.91934, accuracy mencapai 0.91954, dan AUC sebesar 0.96900. Pendekatan balanced split menghasilkan kinerja evaluasi yang lebih efektif dalam memprediksi kelas seimbang, mampu memanfaatkan pola temporal, serta menjadikan model kombinasi lebih unggul dibandingkan model XGBoost.

Item Type: Thesis (Bachelor)
Subjects: T Technology > T Technology (General)
Divisions: Jurusan Matematika dan Teknologi Informasi > Informatika
Depositing User: Sahda Aryanti
Date Deposited: 08 Jul 2025 03:03
Last Modified: 08 Jul 2025 03:03
URI: http://repository.itk.ac.id/id/eprint/22908

Actions (login required)

View Item View Item