Tentang SmartLib Bantuan Pengguna

Skripsi S1

Record Data Skripsi S1

MhswID :

220401156

Judul :

OPTIMASI LONG SHORT TERM MEMORY DENGAN ADAM PADA PREDIKSI NILAI TUKAR RUPIAH TERHADAP DOLAR AS

Penulis :

PAULA CARNELIAN TOBING

Abstrak :

Fluktuasi nilai tukar Rupiah terhadap Dolar Amerika Serikat (USD) merupakan fenomena ekonomi yang berdampak signifikan terhadap stabilitas perekonomian nasional, termasuk sektor perdagangan, investasi, dan kebijakan moneter. Karakteristik data nilai tukar yang bersifat time series dengan pola nonlinier, tren, dan volatilitas tinggi menjadikan proses prediksi sebagai tantangan tersendiri. Penelitian ini bertujuan untuk mengoptimasi algoritma Long Short-Term Memory (LSTM) menggunakan optimizer Adaptive Moment Estimation (Adam) dalam memprediksi nilai tukar Rupiah terhadap Dolar AS. Data historis USD/IDR periode 2010–2025 diperoleh dari Yahoo Finance dengan variabel High dan Low harian. Data diproses menggunakan normalisasi Min–Max Scaling dan dibagi menjadi data latih (80%) dan data uji (20%) dengan skema time step 30 hari. Model dilatih menggunakan Adam dengan learning rate 0,001 dan epoch dengan interval 25 ( 25, 50, 75, 100, 125, 150, 175, 200, 225, 250). Evaluasi dilakukan menggunakan metrik Mean Squared Error (MSE), Root Mean Squared Error (RMSE), dan Mean Absolute Percentage Error (MAPE). Hasil penelitian menunjukkan rata-rata MSE sebesar 37567.76, RMSE sebesar 160.08, dan MAPE sebesar 1.48 %. Nilai MAPE yang sangat rendah serta standar deviasi yang kecil mengindikasikan bahwa model memiliki akurasi tinggi dan performa yang stabil. Dengan demikian, optimasi Adam terbukti efektif dalam meningkatkan kinerja LSTM untuk prediksi nilai tukar Rupiah terhadap Dolar AS. Kata kunci: Deep Learning, LSTM, Nilai Tukar Rupiah, Optimasi Adam, Prediksi Time Series

Abstrak Inggris :

The fluctuation of the Indonesian Rupiah exchange rate against the United States Dollar (USD) is an economic phenomenon that has a significant impact on national economic stability, including the trade sector, investment activities, and monetary policy. The characteristics of exchange rate data as a time series with nonlinear patterns, trends, and high volatility make the prediction process a challenging task. This study aims to optimize the Long Short-Term Memory (LSTM) algorithm using the Adaptive Moment Estimation (Adam) optimizer to predict the Rupiah exchange rate against the US Dollar. Historical USD/IDR data from the period 2010–2025 were obtained from Yahoo Finance using the daily High and Low variables. The data were processed using Min–Max Scaling normalization and divided into training data (80%) and testing data (20%) with a time step scheme of 30 days. The model was trained using the Adam optimizer with a learning rate of 0.001 and epochs with intervals of 25 (25, 50, 75, 100, 125, 150, 175, 200, 225, and 250). Model evaluation was conducted using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). The results show an average MSE of 37,567.76, an RMSE of 160.08, and a MAPE of 1.48%. The relatively low MAPE value and the small standard deviation indicate that the model achieves good accuracy and stable performance. Therefore, the Adam optimization method is considered effective in improving the performance of the LSTM model for predicting the Rupiah exchange rate against the US Dollar. Keywords: Exchange Rate, LSTM, Adam Optimizer, Time Series Prediction, Deep Learning

Tahun :

2026

Cover :

e3a2a-cover-skripsi.pdf

Bab 1 :

Bab 2 :

Bab 3 :

Bab 4 :