Peramalan Outflow Dan Inflow Untuk Transaksi Perbankan Di Bank Indonesia Cabang Balikpapan Menggunakan Model ARIMAX (Variasi Kalender) Dan Long Short-Term Memory (LSTM)

Ramadani, Farham (2025) Peramalan Outflow Dan Inflow Untuk Transaksi Perbankan Di Bank Indonesia Cabang Balikpapan Menggunakan Model ARIMAX (Variasi Kalender) Dan Long Short-Term Memory (LSTM). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003211165-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Bank Indonesia memiliki peran strategis dalam mengelola peredaran uang kartal, khususnya melalui pemantauan arus keluar (outflow) dan arus masuk (inflow) di cabang-cabangnya, termasuk di Balikpapan. Penelitian ini membandingkan dua metode peramalan, yaitu ARIMAX yang mempertimbangkan variasi kalender dan Long Short-Term Memory (LSTM) yang mampu menangkap pola non-linear serta ketergantungan temporal jangka panjang, dalam memprediksi Outflow dan Inflow uang kartal. Analisis dilakukan pada data bulanan periode 2014–2024 dengan memasukkan variabel eksternal seperti bulan Ramadhan, Hari Raya Idul Fitri, dan dampak pandemi COVID-19. Hasil evaluasi menunjukkan bahwa model LSTM unggul dengan nilai RMSE sebesar 209.369 dan MAPE 21,85% untuk Outflow, serta RMSE 99.255 dan MAPE 19,73% untuk Inflow. Sementara ARIMAX menghasilkan RMSE 336.390 dan MAPE 56,1% untuk Outflow, serta RMSE 146.131 dan MAPE 27,7% untuk Inflow. Keunggulan LSTM terutama terlihat pada kemampuannya menangani fluktuasi musiman dan pola non-linear yang kompleks. Penelitian ini mengindikasikan bahwa LSTM lebih sesuai untuk prediksi jangka pendek dan perencanaan pengelolaan uang kartal yang akurat di Bank Indonesia Cabang Balikpapan. Temuan ini memberikan kontribusi metodologis dalam pemodelan data keuangan regional dan menjadi referensi penting dalam mengoptimalkan pengelolaan peredaran uang kartal di tengah dinamika ekonomi yang berkembang, termasuk rencana pemindahan ibu kota negara ke Kalimantan Timur.
===========================================================================================================================================
Bank Indonesia plays a strategic role in managing currency circulation, particularly by monitoring cash outflow and inflow at its branches, including Balikpapan. This study compares two forecasting methods: ARIMAX, which accounts for calendar variations, and Long Short-Term Memory (LSTM), which can capture non-linear patterns and long-term temporal dependencies, to predict cash outflow and inflow. The analysis uses monthly data from 2014 to 2024, incorporating external variables such as Ramadan, Eid al-Fitr, and the COVID-19 pandemic impact. Evaluation results demonstrate that the LSTM model outperforms ARIMAX, achieving an RMSE of 209.369 and MAPE of 21,85% for outflow, and an RMSE of 99.255 and MAPE of 19,73% for inflow. In contrast, ARIMAX yields an RMSE of 336,390 and MAPE of 56.1% for outflow, and an RMSE of 146,131 and MAPE of 27.7% for inflow. LSTM’s superior performance is mainly attributed to its ability to handle seasonal fluctuations and complex non-linear patterns. This study indicates that LSTM is more suitable for short-term forecasting and accurate cash management planning at Bank Indonesia Balikpapan branch. The findings contribute methodologically to regional financial data modeling and provide important insights for optimizing currency circulation management amid evolving economic dynamics, including the planned relocation of the capital city to East Kalimantan.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Outflow, Inflow, ARIMAX, LSTM, Variasi Kalender, Bank Indonesia, Outflow, Inflow, ARIMAX, LSTM, Calendar Variation, Bank Indonesia
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HJ Public Finance Q Science > Q Science (General) > Q325.78 Back propagation Q Science > QA Mathematics > QA280 Box-Jenkins forecasting Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Farham Ramadani
Date Deposited:	31 Jul 2025 08:49
Last Modified:	31 Jul 2025 08:49
URI:	http://repository.its.ac.id/id/eprint/124509

Actions (login required)

View Item