ARSITEKTUR MULTI-TASK MAMBA-2 DENGAN CROSS-DOMAIN GATING MECHANISM BERBASIS MAKROEKONOMI DXY UNTUK PREDIKSI ARAH DAN MAGNITUDO BITCOIN

BAFANA, FIRZAN FAIZ (2026) ARSITEKTUR MULTI-TASK MAMBA-2 DENGAN CROSS-DOMAIN GATING MECHANISM BERBASIS MAKROEKONOMI DXY UNTUK PREDIKSI ARAH DAN MAGNITUDO BITCOIN. S1 thesis, Universitas Mercu Buana Jakarta.

Preview	Text (HAL COVER) Cover.pdf Download (669kB) | Preview
	Text (BAB I) Bab 1.pdf Restricted to Registered users only Download (213kB)
	Text (BAB II) Bab 2.pdf Restricted to Registered users only Download (451kB)
	Text (BAB III) Bab 3.pdf Restricted to Registered users only Download (269kB)
	Text (BAB IV) Bab 4.pdf Restricted to Registered users only Download (750kB)
	Text (BAB V) Bab 5.pdf Restricted to Registered users only Download (238kB)
	Text (DAFTAR PUSTAKA) Daftar Pustaka.pdf Restricted to Registered users only Download (199kB)
	Text (LAMPIRAN) Lampiran.pdf Restricted to Registered users only Download (371kB)

Abstract

The limitations of conventional prediction models in comprehensively capturing cryptocurrency market dynamics drive the need for a more adaptive and integrated modeling framework incorporating global macroeconomic factors. This research aims to design and evaluate a Multi-Task Mamba-2 architecture equipped with a Cross-Domain Gating Mechanism based on the U.S. Dollar Index (DXY) macroeconomic variable to simultaneously predict the directional movement and return magnitude of Bitcoin prices, while benchmarking its performance against a conventional Long Short-Term Memory (LSTM) baseline. The variables examined comprise 116 input features consisting of historical price data, technical indicators, DXY data, and cross-domain fusion features. Data were collected using a totalsampling technique over the period from January 1, 2025 to April 30, 2026 at an hourly resolution (H1), yielding 11,640 raw observations transformed into sliding-window sequences through a chronological split (70% training, 15% validation, and 15% testing). Data analysis was conducted by training five independent random seeds, aggregating a Deep Ensemble, performing statistical significance testing using paired

Item Type:	Thesis (S1)
NIM/NIDN Creators:	41822010097
Uncontrolled Keywords:	Mamba-2, Multi-Task Learning, Cross-Domain Gating, Bitcoin, U.S. DollarIndex (DXY)
Subjects:	000 Computer Science, Information and General Works/Ilmu Komputer, Informasi, dan Karya Umum > 000. Computer Science, Information and General Works/Ilmu Komputer, Informasi, dan Karya Umum > 006 Special Computer Methods/Metode Komputer Tertentu > 006.3 Artificial Intelligence/Kecerdasan Buatan > 006.31 Machine Learning/Pembelajaran Mesin 200 Religion/Agama > 240 Christian Moral and Devotional Theology/Moral Kristen dan Teologi Kebaktian > 246 Use of Art in Christianity/Seni dalam Agama Kristen > 246.9 Architecture/Arsitektur 300 Social Science/Ilmu-ilmu Sosial > 380 Commerce, Communications, Transportation (Perdagangan, Komunikasi, Transportasi) > 381 Commerce, Trade/Perdagangan > 381.1 Retail Trade/Perdagangan Ritail, Pasar
Divisions:	Fakultas Ilmu Komputer > Sistem Informasi
Depositing User:	khalimah
Date Deposited:	23 Sep 2026 04:20
Last Modified:	23 Sep 2026 04:20
URI:	http://repository.mercubuana.ac.id/id/eprint/104065

Actions (login required)

View Item