Paper: arXiv 2403.03606
Authors: Mohammad Ali Labbaf Khaniki, Mohammad Manthouri
Abstract
This study presents an innovative approach for predicting cryptocurrency time series, specifically focusing on Bitcoin, Ethereum, and Litecoin. The methodology integrates the use of technical indicators, a Performer neural network, and BiLSTM (Bidirectional Long Short-Term Memory) to capture temporal dynamics and extract significant features from raw cryptocurrency data. The application of technical indicators, such facilitates the extraction of intricate patterns, momentum, volatility, and trends. The Performer neural network, employing Fast Attention Via positive Orthogonal Random features (FAVOR+), has demonstrated superior computational efficiency and scalability compared to the traditional Multi-head attention mechanism in Transformer models. Additionally, the integration of BiLSTM in the feedforward network enhances the model’s capacity to capture temporal dynamics in the data, processing it in both forward and backward directions. This is particularly advantageous for time series data where past and future data points can influence the current state. The proposed method has been applied to the hourly and daily timeframes of the major cryptocurrencies and its performance has been benchmarked against other methods documented in the literature. The results underscore the potential of the proposed method to outperform existing models, marking a significant progression in the field of cryptocurrency price prediction.
Complexity vs Empirical Score
- Math Complexity: 7.5/10
- Empirical Rigor: 6.5/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced deep learning architectures (Performer, BiLSTM) and technical indicators, demonstrating high mathematical complexity. It includes empirical validation on specific cryptocurrency datasets with comparative benchmarking, indicating substantial backtest-ready implementation.
Research Flowchart
flowchart TD A["Research Goal: Predict Crypto Prices<br>(BTC, ETH, LTC)"] --> B["Data Preparation<br>Hourly & Daily Timeframes + Technical Indicators"] B --> C["Model Architecture: Transformer + BiLSTM"] C --> D["Feature Extraction<br>Performer Network<br>(FAVOR+ Attention)"] D --> E["Temporal Dynamics<br>BiLSTM Layer<br>Forward & Backward Processing"] E --> F["Price Prediction<br>Output Layer"] F --> G["Outcome: Outperforms<br>State-of-the-Art Models"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a new methodology that combines technical indicators, a Performer neural network with FAVOR+ attention, and BiLSTM to predict Bitcoin, Ethereum, and Litecoin prices.
Source sentence
This research introduces a pioneering methodology for time series prediction of Bitcoin, Ethereum, and Litecoin. The approach initially utilizes technical indicators to extract statistical features from the data. Following this, a Performer neural network is applied, which uses Fast Attention Via positive Orthogonal Ra
Assumes
No explicit assumptions found in the text.
Data
- Cryptocurrency price data market-prices unknown
Source sentence
This model is specifically designed to predict the price fluctuations of Bitcoin, Ethereum, and Litecoin span from January 2018 to March 2024 across various time intervals, including daily and hourly.
Costs
Not applicable
Finds 2
- The proposed Transformer-based neural network outperforms other deep learning methods in RMSE for both hourly and daily BTCUSD price prediction. performance
Source sentence
Based on Figures 5 and 6, it can be inferred that the proposed Transformer-based neural network outperforms other deep learning methods.
- Incorporating technical indicators improves the model’s ability to capture market trends and leads to better prediction performance. predictive-accuracy
Source sentence
Moreover, technical indicators provide valuable information about market trends and patterns. When used as feature extractors, they can enhance the model’s ability to capture and learn from these patterns, leading to improved performance.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 8 related papers). Coverage is partial, so this is neither replication nor refutation.
1 paper in the archive
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.