Penerapan Gaussian Process Regression Pada Prediksi Harga Emas Berdasarkan Faktor Makroekonomi

Mardlatillah, Nafisah (2025) Penerapan Gaussian Process Regression Pada Prediksi Harga Emas Berdasarkan Faktor Makroekonomi. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002211037-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Emas merupakan instrumen investasi yang banyak diminati karena kestabilannya dalam menghadapi tekanan ekonomi. Namun, harga emas bersifat fluktuatif dan dipengaruhi oleh berbagai faktor, baik domestik seperti inflasi, suku bunga, dan nilai tukar, maupun global seperti harga emas dunia. Penelitian ini bertujuan untuk memprediksi harga emas ANTAM menggunakan model Gaussian Process Regression (GPR) dengan mempertimbangkan variabel makroekonomi dan harga emas dunia. Data yang digunakan mencakup periode Januari 2020 hingga Januari 2025. Model GPR diatur dengan empat jenis kernel yaitu RBF, Matern, Rational Quadratic, dan Dot Product dengan dua skenario historis, yaitu fitur lag 7 dan lag 30 hari. Hasil analisis korelasi menunjukkan bahwa harga emas dunia memiliki pengaruh paling signifikan terhadap harga emas ANTAM, dengan korelasi sebesar 0,97 (Pearson) dan 0,90 (Spearman), sedangkan variabel lainnya menunjukkan korelasi yang lebih rendah. Hasil evaluasi menunjukkan bahwa kombinasi terbaik diperoleh dari kernel Dot Product dengan lag 7, yang menghasilkan nilai MAPE sebesar 0,23% dan RMSE 3704,79 pada tahap validasi, serta MAPE 0,39% dan RMSE 7965,63 pada tahap pengujian. Dengan demikian, pendekatan GPR dengan kernel Dot Product dan lag pendek terbukti paling efektif dan stabil dalam menangkap pola dan memprediksi harga emas ANTAM.
=========================================================================================================================================
Gold is a popular investment instrument due to its stability in facing economic pressures. However, gold prices are volatile and influenced by various factors, both domestic such as inflation, interest rates, and exchange rates—and global, such as the world gold price. This study aims to predict ANTAM gold prices using the Gaussian Process Regression (GPR) model by considering macroeconomic variables and the world gold price. The data used covers the period from January 2020 to January 2025. The GPR model is configured with four types of kernels, namely RBF, Matern, Rational Quadratic, and Dot Product, under two historical scenarios, namely 7-day and 30-day lag features. Correlation analysis shows that the world gold price has the most significant influence on ANTAM gold prices, with a correlation of 0.97 (Pearson) and 0.90 (Spearman), while the other variables show lower correlations. The evaluation results indicate that the best combination is obtained from the Dot Product kernel with a 7-day lag, producing a MAPE of 0.23% and RMSE of 3704.79 in the validation stage, and a MAPE of 0.39% and RMSE of 7965.63 in the testing stage. Thus, the GPR approach with the Dot Product kernel and short lag proves to be the most effective and stable in capturing patterns and predicting ANTAM gold prices.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Prediksi, Emas, ANTAM, Gaussian Process Regression (GPR), Makroekonomi
Subjects:	Q Science > QA Mathematics > QA246.8 Gaussian Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA353.K47 Kernel functions (analysis)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Nafisah Mardlatillah
Date Deposited:	04 Aug 2025 03:49
Last Modified:	04 Aug 2025 03:49
URI:	http://repository.its.ac.id/id/eprint/126811

Actions (login required)

View Item