IMPLEMENTASI KINERJA ALGORITMA MACHINE LEARNING DALAM PREDIKSI HARGA BITCOIN,ETHEREUM DAN USD:STUDI KASUS YAHOO FINANCE STOCK MARKET

Ihsannur Fathan Riziq; Achmad Rizal Dzikrillah;

ABSTRAK

Implementasi Kinerja Algoritma Machine Learning Dalam Prediksi Harga Bitcoin,Ethereum Dan Usd:Studi Kasus Yahoo Finance Stock Market

Ihsannur Fathan Riziq

Pengembangan teknologi di sektor keuangan telah memfasilitasi lahirnya berbagai instrumen
investasi digital, salah satunya cryptocurrency. Bitcoin dan Ethereum merupakan aset digital
dengan kapitalisasi pasar terbesar, sedangkan USD tetap memegang peran penting dalam
perdagangan global. Tingginya volatilitas harga dari ketiga aset ini menuntut metode prediksi
yang akurat dan adaptif. Penelitian ini bertujuan untuk menerapkan algoritma machine learning
Long Short-Term Memory (LSTM) dalam memprediksi harga Bitcoin, Ethereum, dan USD
berdasarkan data historis dari Yahoo Finance periode 2019 hingga 2024. Proses prapemrosesan
meliputi normalisasi data dengan Min-Max Scaler dan rekayasa fitur berupa return harian.
Evaluasi model dilakukan menggunakan metrik Mean Absolute Error (MAE) dan Mean Squared
Error (MSE). Hasil penelitian menunjukkan bahwa model LSTM memiliki performa terbaik,
dengan nilai MAE terendah sebesar 1.320,41 dan MSE sebesar 3.464.596,53 pada prediksi harga
tertinggi (high). Temuan ini menunjukkan bahwa LSTM unggul dalam menangani pola data
kompleks dan fluktuatif secara konsisten. Penelitian ini diharapkan dapat menjadi acuan dalam
pengembangan sistem prediksi harga aset digital berbasis machine learning, khususnya untuk
aset dengan volatilitas tinggi.

Kata Kunci : Machine Learning,Cryptocurrency,USD,SVR,LSTM

Impelementation of Machine Learning Alogrithms Perfomance In Predicting Bitcoin,Ethereum, And Usd Prices : A Case Study on Yahoo Finance Stock Market

Ihsannur Fathan Riziq

Technological developments in the financial sector have facilitated the emergence of various digital
investment instruments, one of which is cryptocurrency. Bitcoin and Ethereum are digital assets with
the largest market capitalization, while the USD remains a significant player in global trade. The high
price volatility of these three assets demands accurate and adaptive prediction methods. This study
aims to apply the Long Short-Term Memory (LSTM) learning algorithm to predict Bitcoin, Ethereum,
and USD prices based on historical data from Yahoo Finance from 2019 to 2024. Preprocessing
includes data normalization with a Min-Max Scaler and feature engineering in the form of daily
returns. Model evaluation was conducted using the Mean Absolute Error (MAE) and Mean Squared
Error (MSE) metrics. The results showed that the LSTM model performed best, with the lowest MAE
value of 1,320.41 and an MSE of 3,464,596.53 for the highest price prediction. These findings
demonstrate that LSTM excels in consistently handling complex and fluctuating data patterns. This
research is expected to serve as a reference in the development of a machine learning-based digital
asset price prediction system, particularly for assets with high volatility.

Keywords: Machine learning; Cryptocurrency; Unite State Dolar; Long Short Term Memory;Support Vector Regression