Sign In

TFT-ACB-XML: Decision-Level Integration of Customized Temporal Fusion Transformer and Attention-BiLSTM with XGBoost Meta-Learner for BTC Price Forecasting

Created by
  • Haebom
Category
Empty

μ €μž

Raiz Ud Din (Artificial Intelligence Lab, Department of Computer Systems Engineering, University of Engineering and Applied Sciences), Saddam Hussain Khan (Interdisciplinary Research Center for Smart Mobility and Logistics, King Fahad University of Petroleum and Minerals)

πŸ’‘ κ°œμš”

λ³Έ μ—°κ΅¬λŠ” λΉ„μ„ ν˜•μ μ΄κ³  변동성이 큰 λΉ„νŠΈμ½”μΈ(BTC) 가격 예츑의 어렀움을 ν•΄κ²°ν•˜κΈ° μœ„ν•΄, Temporal Fusion Transformer (TFT)와 Attention-BiLSTM (ACB) 두 개의 병렬 기반 ν•™μŠ΅κΈ°λ₯Ό XGBoost 메타 ν•™μŠ΅κΈ°μ™€ κ²°ν•©ν•œ ν•˜μ΄λΈŒλ¦¬λ“œ μŠ€νƒ μΌλ°˜ν™” ν”„λ ˆμž„μ›Œν¬μΈ TFT-ACB-XML을 μ œμ•ˆν•©λ‹ˆλ‹€. 이 ν”„λ ˆμž„μ›Œν¬λŠ” μž₯기적/단기적 μ‹œκ³„μ—΄ μ˜μ‘΄μ„±μ„ 효과적으둜 ν¬μ°©ν•˜κ³ , 검증 μ„±λŠ₯ 기반의 κ°€μ€‘μΉ˜ λΆ€μ—¬λ₯Ό 톡해 예츑 정확도λ₯Ό ν–₯μƒμ‹œν‚΅λ‹ˆλ‹€.

πŸ”‘ μ‹œμ‚¬μ  및 ν•œκ³„

β€’
μ‹œμ‚¬μ  1: μ‹œκ³„μ—΄ λ°μ΄ν„°μ˜ μž₯κΈ° 및 단기 μ˜μ‘΄μ„±μ„ λ™μ‹œμ— κ³ λ €ν•˜λŠ” 앙상블 기법이 λΉ„νŠΈμ½”μΈκ³Ό 같이 λ³΅μž‘ν•œ 금육 μ‹œμž₯ μ˜ˆμΈ‘μ—μ„œ 단일 λͺ¨λΈλ³΄λ‹€ μš°μˆ˜ν•œ μ„±λŠ₯을 보일 수 μžˆμŒμ„ μž…μ¦ν–ˆμŠ΅λ‹ˆλ‹€.
β€’
μ‹œμ‚¬μ  2: 해석 κ°€λŠ₯ν•œ 주의 λ©”μ»€λ‹ˆμ¦˜μ„ ν™œμš©ν•˜μ—¬ λͺ¨λΈμ˜ μ„€λͺ…λ ₯을 λ†’μ΄λ©΄μ„œλ„, XGBoostλ₯Ό 톡해 λΉ„μ„ ν˜•μ μΈ μž”μ°¨κΉŒμ§€ ν•™μŠ΅ν•¨μœΌλ‘œμ¨ 예츑 μ„±λŠ₯을 κ·ΉλŒ€ν™”ν•˜λŠ” 효과적인 ν•˜μ΄λΈŒλ¦¬λ“œ μ ‘κ·Ό 방식을 μ œμ‹œν–ˆμŠ΅λ‹ˆλ‹€.
β€’
ν•œκ³„μ  λ˜λŠ” ν–₯ν›„ 과제: μ œμ•ˆλœ λͺ¨λΈμ˜ μ „λ°˜μ μΈ μ„±λŠ₯은 μš°μˆ˜ν•˜λ‚˜, νŠΉμ • κΈ°κ°„μ΄λ‚˜ μ‹œμž₯ μƒν™©μ—μ„œ λ‹€λ₯Έ 앙상블 μ‘°ν•© λ˜λŠ” κ³ κΈ‰ 메타 ν•™μŠ΅κΈ°κ°€ 좔가적인 κ°œμ„ μ„ κ°€μ Έμ˜¬ 수 μžˆλŠ”μ§€μ— λŒ€ν•œ 좔가적인 탐색이 ν•„μš”ν•©λ‹ˆλ‹€. λ˜ν•œ, λ‹€μ–‘ν•œ μ•”ν˜Έν™”ν μ‹œμž₯으둜의 μΌλ°˜ν™” κ°€λŠ₯μ„± 및 μ‹€μ‹œκ°„ 예츑 μ‹œμŠ€ν…œμ—μ„œμ˜ 적용 κ°€λŠ₯성에 λŒ€ν•œ 검증이 μš”κ΅¬λ©λ‹ˆλ‹€.
πŸ‘
Made with Slashpage