Peramalan Nilai Tukar Rupiah Terhadap USD Menggunakan Metode Hybrid Autoregressive Integrated Moving Average-Long Short Term Memory (ARIMA-LSTM)

Respati, Yosafat (2024) Peramalan Nilai Tukar Rupiah Terhadap USD Menggunakan Metode Hybrid Autoregressive Integrated Moving Average-Long Short Term Memory (ARIMA-LSTM). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006201029-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (3MB) | Request a copy

Abstract

Pasar valuta asing merupakan tempat jual beli mata uang suatu negara dengan negara lain. Mata uang asing diperdagangkan berdasarkan nilai tukar tertentu. Sistem nilai tukar pada pasar valuta asing menggunakan sistem free-floating, dimana nilai tukar mata uang dapat berubah-ubah berdasarkan jumlah penawaran dan permintaan. Nilai tukar mata uang yang berubah-ubah dapat dimanfaatkan oleh investor untuk melakukan investasi, seperti spot, futures, dan options. Namun, hal tersebut bisa menjadi risiko bagi perusahaan multinasional atau orang yang melakukan jual beli internasional karena dapat memengaruhi jumlah keuntungan dari transaksi bisnis. Mengetahui pergerakan nilai tukar mata uang merupakan hal krusial karena dapat mempengaruhi keputusan investasi dan perencanaan keuangan perusahaan jangka pendek. Pada penelitian ini dilakukan analisis time series nilai tukar mata uang Rupiah terhadap US Dollar menggunakan metode hybrid Autoregressive Integrated Moving Average-Long Short Term Memory (ARIMA-LSTM). Variabel yang digunakan dalam penelitian ini adalah data historis sekunder nilai tukar mata uang Rupiah terhadap Dollar Amerika yang diperoleh melalui situs Bank Indonesia pada periode 4 Januari 2010 – 31 Agustus 2023. Pada periode tersebut karakteristik nilai kurs cenderung naik yang menandakan bahwa Rupiah mengalami pelemahan selama periode tersebut. Melalui penelitian ini, didapatkan hasil model terbaik yaitu ARIMA (2,1,0)(3,0,1)5 dengan LSTM (4-3-1) dengan input lag signifikan y_(t-1), y_(t-2), y_(t-12), dan y_(t-14). Model tersebut mendapatkan nilai evaluasi sebesar 4,74962%. Model hybrid ARIMA (2,1,0)(3,0,1)5-LSTM (4-3-1) menurunkan nilai error sebesar 0,031% dari model ARIMA. Menggunakan model terbaik tersebut, didapatkan hasil peramalan nilai tukar Rupiah terhadap USD cenderung stabil di kisaran 15.200. Penelitian ini diharapkan dapat membantu investor dalam pengambilan keputusan investasi dan perusahaan multinasional dalam merencanakan keuangan jangka pendeknya.
=================================================================================================================================
Foreign exchange market is a place to buy and sell currency from one country to other country currency. Foreign currency is sold based on exchange rate value. Exchange rates use free-float system that make exchange rate fluctuate according to supply and demand on the market. This fluctuation is used by investors to invest in foreign exchange. They invest in spot trading, futures, and options. However, this fluctuation is also a risk for multinational companies or someone who trades internationally because it affects their profit margin. Understanding how exchange rate movement is crucial because it can influence investment decision and short-term financing planning. This study will analyze Indonesian Rupiah to American Dollar exchange rate using hybrid Autoregressive Integrated Moving Average-Long Short Term Memory (ARIMA-LSTM) method. The variable used in this study is Indonesian Rupiah to American Dollar exchange rate historical secondary data that collected from Bank Indonesia in 4 January 2010 – 31 August 2023. In that period, the characteristics of the exchange rate tended to increase, indicating that the Rupiah weakened. Through this research, the best model is ARIMA (2,1,0)(3,0,1)5 with LSTM (4-3-1) with significant lag inputs y_(t-1), y_(t-2), y_(t-12), and y_(t-14). The model obtained an evaluation value of 4.74962%. The hybrid model reduces the error value by 0.031% compared to the ARIMA model. Using this model, the results of forecasting Indonesian Rupiah exchange rate against the US Dollar tend to stabilize around 15,200. This research is expected to help investors in making investment decisions and multinational companies in planning their short-term finances.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	ARIMA, Artificial Neural Network, Foreign Exchange, Jaringan Saraf Tiruan, LSTM, Time Series, Valuta Asing.
Subjects:	Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Yosafat Respati
Date Deposited:	31 Jan 2024 02:08
Last Modified:	31 Jan 2024 02:08
URI:	http://repository.its.ac.id/id/eprint/105776

Actions (login required)

View Item