Rancang Bangun Dashboard Prediksi Harga Bitcoin dengan Menggunakan Long Short Term Memory (LSTM) Memory

Farihah, Tatik Farihatul (2024) Rancang Bangun Dashboard Prediksi Harga Bitcoin dengan Menggunakan Long Short Term Memory (LSTM) Memory. Diploma thesis, Institut Teknologi Sepuluh Nopember.

Text 2043211023-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (3MB) | Request a copy

Abstract

Era teknologi saat ini, investasi pada cryptocurrency telah tumbuh signifikan, terutama di Indonesia yang menduduki peringkat ketujuh dunia pada Maret 2024. Bitcoin, sebagai cryptocurrency dominan, menunjukkan volatilitas harga yang tinggi, menawarkan peluang investasi besar namun dengan risiko tinggi. Oleh karena itu, diperlukan alat prediksi efektif untuk membantu investor dalam pengambilan keputusan. Penelitian ini berfokus pada pemodelan prediksi harga Bitcoin menggunakan Long Short Term Memory (LSTM) Network dari Januari 2023 hingga Juli 2024. Tiga skenario model prediksi dibandingkan: model berdasarkan data historis harga Bitcoin, model dengan variabel internal dan eksternal, serta model yang mengombinasikan kedua jenis variabel tersebut. Evaluasi performa model dilakukan menggunakan Mean Absolute Percentage Error (MAPE) untuk menemukan model terbaik. Penelitian ini diharapkan membantu investor dalam pengambilan keputusan yang lebih informasi, mendukung pengawasan pasar oleh BAPPEBTI, dan memberikan pemahaman lebih baik tentang dinamika pasar cryptocurrency bagi pemerintah. Penelitian ini juga mencakup pengembangan dashboard prediksi harga Bitcoin yang interaktif, memanfaatkan kecerdasan perangkat lunak. Data penelitian mencakup harga Bitcoin, volume perdagangan, Nasdaq, dan S&P 500. Langkah analisis meliputi pembersihan data, analisis pola harga, pemodelan LSTM, dan evaluasi performa model.
=================================================================================================================================
In the current technological era, cryptocurrency investment has grown significantly, especially in Indonesia, which ranked seventh globally in March 2024. Bitcoin, as the dominant cryptocurrency, exhibits high price volatility, offering substantial investment opportunities but with high risk. Therefore, an effective prediction tool is necessary to assist investors in decision-making. This study focuses on modeling Bitcoin price predictions using the Long Short Term Memory (LSTM) Network from January 2023 to July 2024. Three prediction model scenarios are compared: a model based on historical Bitcoin price data, a model incorporating internal and external variables, and a model combining both types of variables. Model performance is evaluated using Mean Absolute Percentage Error (MAPE) to identify the best model. This research aims to assist investors in making more informed decisions, support market supervision by BAPPEBTI, and provide the government with a better understanding of cryptocurrency market dynamics. The study also includes the development of an interactive Bitcoin price prediction dashboard, leveraging artificial intelligence. Data for this study encompassing Bitcoin prices, trading volume, Nasdaq, and S&P 500. The analysis steps include data cleaning, price pattern analysis, LSTM modeling, and model performance evaluation.

Item Type:	Thesis (Diploma)
Uncontrolled Keywords:	Bitcoin, Long Short Term Memory Network, Jaringan Memori Jangka Panjang Pendek
Subjects:	T Technology > T Technology (General) T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Tatik Farihatul Farihah
Date Deposited:	19 Dec 2024 08:43
Last Modified:	19 Dec 2024 08:43
URI:	http://repository.its.ac.id/id/eprint/116019

Actions (login required)

View Item