An empirical evaluation of fuzzy bidirectional long short-term memory with soft computing based decision-making model for predicting volatility of cryptocurrencies
Evaluates a hybrid FBLSTM-ICPA model integrating fuzzy logic, normalization, and hyperparameter tuning to predict cryptocurrency volatility.
What it examines
This paper introduces a hybrid model, FBLSTMSC-DMPVC, that combines fuzzy logic with Bidirectional LSTM optimized by an improved Carnivorous Plant Algorithm. It uses Z-score normalization to standardize data, aiming to accurately predict cryptocurrency volatility and support decision-making in dynamic financial markets.
What it concludes
The model achieves superior performance with low error metrics across BTC, ETH, LTC, and XRP, enhancing volatility prediction. It supports investor and policymaker decisions, indicating applications in portfolio management and risk assessment. Future studies may refine the approach and explore broader uses in financial market forecasting.
Evidence objects
Researchers introduce the novel FBLSTMSC-DMPVC framework, integrating fuzzy logic, bidirectional LSTM, and an improved carnivorous plant algorithm, capturing past and future dependencies in cryptocurrency volatility forecasting with low error rates.
key_findings bullet 1 · key_findings · validation V0
The study employs comprehensive Z-score normalization, simulation evaluations on 2018-2023 historical data, and rigorous comparisons with established models, highlighting significant improvements using MAPE values as low as $$0.6187$$ and $$0.6667$$.
key_findings bullet 2 · key_findings · validation V0
Authors introduce new fuzzy mapping and implication rule terminologies while addressing non-linear cryptocurrency behaviors; however, reliance on simulations raises concerns over computational complexity, overfitting risks, and challenges for real-time implementation.
key_findings bullet 3 · key_findings · validation V0
The paper introduces an innovative hybrid approach that combines fuzzy logic with bidirectional LSTM and an improved hyperparameter tuning algorithm ($\text{ICPA}$) for forecasting cryptocurrency volatility. Its original integration of soft computing and deep learning within DeFi and quantitative finance offers a novel, compelling perspective that promises significant impact and insights.
key_findings bullet 4 · key_findings · validation V0
Raw abstract and provenance
- … With the growing use of AI in various fields, its application in financial markets, … as Deep Learning (DL) and Machine Learning (ML) methods, are utilized to forecast time …
Source row: 180 · abstract type: snippet