Peramalan Trafik Layanan Data (Payload) Operator X di Area 3 Jawa Bali Menggunakan ARIMAX dan Long Short-Term Memory

Pasaribu, Naomi Gloria (2026) Peramalan Trafik Layanan Data (Payload) Operator X di Area 3 Jawa Bali Menggunakan ARIMAX dan Long Short-Term Memory. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221016-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Pertumbuhan kebutuhan masyarakat terhadap layanan data digital di Indonesia terus meningkat. Operator X sebagai perusahaan telekomunikasi terbesar di Indonesia mengalami fluktuasi trafik layanan data (payload) pada momen tertentu seperti Hari Raya Natal dan Tahun Baru, Idul Fitri, serta penurunan drastis saat Nyepi akibat penghentian sementara layanan internet seluler di wilayah Bali. Kondisi ini menimbulkan tantangan dalam pengelolaan trafik data yang adaptif terhadap dinamika musiman dan variasi kalender terutama di wilayah operasional Area 3 Jawa Bali sebagai kawasan dengan trafik data tinggi. Penelitian ini bertujuan untuk membandingkan dua pendekatan peramalan yaitu ARIMAX dan LSTM untuk menentukan metode yang paling efektif dalam menangkap pola musiman serta dinamika trafik layanan data (payload). Hasil analisis menunjukkan bahwa trafik layanan data (payload) mengalami tren peningkatan yang stabil dengan pola musiman mingguan dan lonjakan pada hari besar keagamaan. Model ARIMAX ([1, 7], 1, [1, 7]) menghasilkan MAE, MAPE, dan RMSE pada data out sample berturut-turut sebesar 0,653; 4,986%; dan 0,789 serta belum mampu menangkap pola nonlinear dan lonjakan ekstrem pada data aktual. Model LSTM dengan konfigurasi terbaik yaitu batch size 16, jumlah neuron 150, hidden layer 1, epoch 100, dan learning rate 0,001 menghasilkan MAE, MAPE, dan RMSE pada data testing berturut-turut sebesar 0,173; 1,32%; dan 0,245 serta mampu menangkap pola nonlinear dan lonjakan ekstrem dengan lebih akurat. Oleh karena itu, model LSTM dipilih sebagai model terbaik untuk meramalkan trafik layanan data (payload) Operator X di Area 3 Jawa Bali periode Juli dan Agustus tahun 2025. Hasil penelitian diharapkan dapat memberikan rekomendasi strategi peramalan berbasis data yang mendukung pengambilan keputusan operasional oleh Operator X.
================================================================================================================================
The growth in public demand for digital data services in Indonesia continues to increase. Operator X, as the largest telecommunications company in the country, experiences fluctuations in data service traffic (payload) during certain events such as Christmas and New Year, Eid al-Fitr, as well as a sharp decline during Nyepi due to the temporary suspension of mobile internet services in the Bali region. These conditions create challenges in managing data traffic that must adapt to seasonal dynamics and calendar variations, particularly in the operational Area 3 of Java Bali, which is characterized by high data usage. This study aims to compare two forecasting approaches, ARIMAX and LSTM, to determine which method is more effective in capturing seasonal patterns and the dynamics of data service traffic (payload). The analysis shows that data traffic exhibits a stable upward trend with weekly seasonality and significant spikes on major religious holidays. The ARIMAX model ([1, 7], 1, [1, 7]) produces MAE, MAPE, and RMSE values on the out-of-sample data of 0.653, 4.986%, and 0.789, respectively, and it has limitations in capturing nonlinear patterns and extreme spikes in the actual data. Meanwhile, the LSTM model with the optimal configuration batch size 16, 150 neurons, 1 hidden layer, 100 epochs, and a learning rate of 0.001 produces MAE, MAPE, and RMSE values of 0.173, 1.32%, and 0.245 on the testing data. This model is able to capture nonlinear patterns and extreme fluctuations more accurately. Therefore, the LSTM model is selected as the best model for forecasting data service traffic (payload) for Operator X in Area 3 of Java Bali for July and August 2025. The results of this study are expected to provide data-driven forecasting recommendations to support operational decision-making by Operator X.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	ARIMAX, LSTM, Operator X, Payload, Trafik, ARIMAX, LSTM, Operator X, Payload, Traffic
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HD Industries. Land use. Labor > HD30.27 Business forecasting Q Science > QA Mathematics > QA280 Box-Jenkins forecasting R Medicine > R Medicine (General) > R858 Deep Learning
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Naomi Gloria Pasaribu
Date Deposited:	25 Feb 2026 01:53
Last Modified:	25 Feb 2026 01:53
URI:	http://repository.its.ac.id/id/eprint/132597

Actions (login required)

View Item