PENERAPAN MODEL GATED RECURRENT UNIT UNTUK PERAMALAN JUMLAH PENUMPANG KERETA API DI PT. KAI (Persero)

WIBOWO, SATRIO BAYU (2022) PENERAPAN MODEL GATED RECURRENT UNIT UNTUK PERAMALAN JUMLAH PENUMPANG KERETA API DI PT. KAI (Persero). S1 thesis, Universitas Mercu Buana Jakarta-Menteng.

Preview	Text (Cover) COVER - SATRIO BAYU WIBOWO.pdf Download (1MB) | Preview
Preview	Text (Abstrak) Abstrak - SATRIO BAYU WIBOWO.pdf Download (382kB) | Preview
	Text (Bab 1) BAB 1 - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (254kB)
	Text (Bab 2) BAB 2 - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (892kB)
	Text (Bab 3) BAB 3 - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (620kB)
	Text (Bab 4) BAB 4 - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (341kB)
	Text (Bab 5) BAB 5 - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (849kB)
	Text (Daftar Pustaka) Daftar Pustaka - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (279kB)
	Text (Lampiran) Lampiran - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (312kB)
	Text (Formulir pernyataan keabsahan dan persetujuan publikasi tugas akhir) Hasil Scan Formulir Pernyataan Keabsahan dan Persetujuan Publikasi Tugas Akhir - SATRIO BAYU WIBOWO.pdf Restricted to Registered users only Download (157kB)

Abstract

Pada penelitian ini pengggunaan data penumpang PT.KAI memiliki rentang waktu setiap bulan dengan jumlah data penumpang sebanyak 169 bulan pada bulan Januari 2006 sampai Januari 2020. Jumlah penumpang PT.KAI dapat di ramal dengan menggunakan model Gated Recurrent Unit (GRU) dengan bahasa pemprograman python. 169 data bulanan ini akan di bagi menjadi 2 data, yaitu data latih sebesar 64% dan data uji sebesar 36%. Untuk mendapatkan performa yang lebih tinggi saat melatih model GRU diberikan inisialisasi Hyperparameter adalah Learning Rate sebesar 0,01, Batch Size sebanyak 100, Hidden State 512, Windows Size 30 dan Epoch sampai 15000. Berdasarkan hasil pengujian didapatkan model yang terbaik pada Epoch ke-14000 yang memiliki loss terkecil sebesar 1.08 × 10−10 . Kemudian model tersebut di ujikan pada data uji dan di dapatkan nilai mean absolute percentage error (MAPE) sebesar 4,84%. Kata kunci: Gated Recurrent Unit, Mean Absolute Percetage Error, Peramalan, PT. Kereta Api Indonesia. In this study, we use PT. KAI passenger data which has a period every month with a total passenger data of 169 months from January 2006 to January 2020. The number of PT. KAI passengers predicted using the Gated Recurrent Unit (GRU) model. From 169 data, it divided into two data, which are training data for 64% and test data for 36%. To get better performance when training the GRU model, hyperparameter initialization given by a learning rate of 0.01, a batch size of 100, hidden state 512, windows size 30, and epoch up to 15000. Based on the test results in a good model on the epoch to 14000 which has the smallest loss. The best model is tested on test data and obtains a mean absolute percentage error (MAPE) value of 4.84 %. Key words: Gated Recurrent Unit, Mean Absolute Percetage Error, forcesting, Indonesian Railways Company.

Item Type:	Thesis (S1)
NIM/NIDN Creators:	41520110032
Uncontrolled Keywords:	Kata kunci: Gated Recurrent Unit, Mean Absolute Percetage Error, Peramalan, PT. Kereta Api Indonesia. Key words: Gated Recurrent Unit, Mean Absolute Percetage Error, forcesting, Indonesian Railways Company.
Subjects:	000 Computer Science, Information and General Works/Ilmu Komputer, Informasi, dan Karya Umum > 000. Computer Science, Information and General Works/Ilmu Komputer, Informasi, dan Karya Umum > 004 Data Processing, Computer Science/Pemrosesan Data, Ilmu Komputer, Teknik Informatika
Divisions:	Fakultas Ilmu Komputer > Informatika
Depositing User:	SITI NOVI NUR CAHYANI
Date Deposited:	07 Jan 2023 04:19
Last Modified:	07 Jan 2023 04:19
URI:	http://repository.mercubuana.ac.id/id/eprint/73154

Actions (login required)

View Item