Dimitris Kastoris, Dimitris Papadopoulos, Konstantinos C. Giotopoulos
Mathematical modeling plays a crucial role in supporting decision-making across a wide range of scientific disciplines. These models often involve multiple parameters, the estimation of which is critical to assessing their reliability and predictive power. Recent advancements in artificial intelligence have made it possible to efficiently estimate such parameters with high accuracy. In this study, we focus on modeling the dynamics of cryptocurrency market shares by employing a Lotka-Volterra system. We introduce a methodology based on a deep neural network (DNN) to estimate the parameters of the Lotka-Volterra model, which are subsequently used to numerically solve the system using a fourth-order Runge-Kutta method. The proposed approach, when applied to real-world market share data for Bitcoin, Ethereum, and alternative cryptocurrencies, demonstrates excellent alignment with empirical observations. Moreover, our method outperforms ARIMA models in terms of accuracy, showcasing its effectiveness for crypto market forecasting. The entire framework, including neural network training and Runge-Kutta integration, was implemented in MATLAB.
Makine öğrenmesi ve derin öğrenme, son yıllarda finansal piyasalarda tahminleme süreçlerinde sıklıkla başvurulan yöntemler arasında yer almaktadır. Bu yöntemler, özellikle yüksek oynaklığa sahip kripto para piyasalarında, yatırım kararlarını destekleyici araçlar olarak öne çıkmaktadır. Denetimli öğrenme algoritmaları, geçmiş verilere dayanarak gelecekteki fiyatları tahmin etmeye yönelik güçlü çözümler sunarken; derin öğrenme yaklaşımları, özellikle sıralı veri yapılarındaki karmaşık ilişkileri yakalamada avantaj sağlamaktadır. Bitcoin gibi dijital varlıkların ekonomik değişkenlerle olan ilişkisini anlamak, hem bireysel yatırımcılar hem de finansal kurumlar açısından stratejik bir gereklilik haline gelmiştir. Bu bağlamda açıklanabilir yapay zeka (XAI) teknikleri, tahmin modellerinin iç mantığını şeffaf biçimde ortaya koyarak karar destek sistemlerine katkı sunmaktadır. Bu çalışmada, Bitcoin’in (BTC) günlük kapanış fiyatları, altın, USDX, VIX ve Brent petrol gibi ekonomik göstergeler dikkate alınarak SVR, LSTM, XGBoost ve ANN algoritmaları ile tahmin edilmiştir. 2014–2024 dönemini kapsayan veri setiyle gerçekleştirilen analizde, MAPE, MAE ve R² gibi performans ölçütleri üzerinden karşılaştırma yapılmıştır. Sonuçlara göre, en yüksek doğruluk oranını SVR modeli göstermiştir. Ayrıca SHAP analizi kullanılarak BTCY (yükseliş), BTCD (düşüş) ve BTCA (açılış) değişkenlerinin tahmin sürecinde pozitif katkı sunduğu belirlenmiştir. Buna karşılık, AK (altın) ve BrPK (petrol) gibi dışsal faktörlerin olumsuz etkiler yarattığı gözlemlenmiştir. Elde edilen bulgular ile hem model doğruluğunu hem de değişkenlerin etkisini şeffaf biçimde ortaya koyarak literatüre önemli katkılar sunulmuştur.
Ethereum price forecasting remains a challenging task due to the highly volatile and nonlinear nature of digital asset markets. This study proposes a hybrid time-series forecasting model that integrates an Autoregressive Integrated Moving Average (ARIMA) model with a Long Short-Term Memory (LSTM) network to improve Ethereum price prediction accuracy. The ARIMA model captures linear dependencies and extracts statistical residuals, which are then incorporated as an additional input feature for the LSTM network to enhance its learning of complex temporal patterns. The model is trained using a dataset containing historical Ethereum price data, with MinMax normalization applied to the closing prices for improved stability and provides 0.398821 MAE. A comprehensive ablation study evaluates different model configurations, demonstrating that the ARIMA residuals significantly enhance predictive performance. The hybrid ARIMA-LSTM model achieves a Mean Squared Error (MSE) of 0.1846, outperforming standalone LSTM and ARIMA models. Further residual and error analysis confirm that the model effectively mitigates autocorrelation in forecasting errors while maintaining stable predictive performance.
In this note, we make a comparison between a novel machine learning method, Long Short-Term Memory (LSTM), and two trading strategies using technical analysis: Exponential Moving Average (EMA) crossing and Moving Average Convergence/Divergence with Average Directional Index (MACD+ADX). The purpose is to use trading signals to maximize profits in the Bitcoin digital commodity. The comparison was motivated by the approval of the first spot Bitcoin exchange-traded funds (ETFs) by the U.S. Securities and Exchange Commission (SEC) on January 9, 2024. The results show that the LSTM algorithm delivers a cumulative return of approximately 65.23% over a testing period of less than nine months, significantly outperforming both the EMA and MACD+ADX strategies, as well as the baseline buy-and-hold approach typically followed by fundamental investors. Our work highlights the potential for further integration between machine learning and technical analysis in the evolving landscape of cryptocurrency markets.
Cryptocurrency price prediction remains a challenging task due to high market volatility, complex price fluctuations, and external influencing factors. Traditional financial models struggle to capture these dynamic patterns, leading to inconsistent forecasting accuracy. This research presents a hybrid deep learning and machine learning approach that integrates convolutional neural networks (CNN) for feature extraction and random forest (RF) for interpretability. By utilizing historical price data from Binance, CoinGecko, and CoinMarketCap, the model leverages both numerical indicators and graphical price representations to improve prediction reliability. The results demonstrate that the hybrid model achieves higher accuracy compared to standalone models. The model records a mean absolute error (MAE) of 0.042 and a root mean square error (RMSE) of 0.078, outperforming both individual CNN and RF models. Feature importance analysis reveals that trading volume, moving averages, and MACD are the most influential factors in price forecasting. Additionally, the model successfully identifies profitable trading signals, achieving an overall prediction accuracy of 89.7 percent. These findings highlight the effectiveness of integrating deep learning with machine learning for cryptocurrency price prediction. The proposed model enhances prediction accuracy and decision-making, making it a valuable tool for traders and financial analysts navigating volatile cryptocurrency markets.
This paper proposes a comprehensive deep-learning framework, SentiStack, for Bitcoin price forecasting and trading strategy evaluation by integrating multimodal data sources, including market indicators, macroeconomic variables, and sentiment information extracted from financial news and social media. The model architecture is based on a Stacking-LSTM ensemble, which captures complex temporal dependencies and non-linear patterns in high-dimensional financial time series. To enhance predictive power, sentiment embeddings derived from full-text analysis using the DeepSeek language model are fused with traditional numerical features through early and late data fusion techniques. Empirical results demonstrate that the proposed model significantly outperforms baseline strategies, including Buy & Hold and Random Trading, in cumulative return and risk-adjusted performances. Feature ablation experiments further reveal the critical role of sentiment and macroeconomic inputs in improving forecasting accuracy. The sentiment-enhanced model also exhibits strong performance in identifying high-return market movements, suggesting its practical value for data-driven investment decision-making. Overall, this study highlights the importance of incorporating soft information, such as investor sentiment, alongside traditional quantitative features in financial forecasting models.
Purpose The aim of this research is to analyze the relationship between the MAX effect and the sentiment associated with the news in the cryptocurrency market. Design/methodology/approach The study uses natural language processing to build the sentiment indicator derived from news headlines about cryptocurrencies. Further, a survey-based sentiment indicator is also utilized. The study undertakes analysis at both the portfolio level (Decile analysis) and at the cross-sectional level (using the Fama-Macbeth Regression). Findings The results demonstrate a positive MAX effect in the cryptocurrency market. When the investor sentiment interacts with the MAX effect, the positive MAX effect continues to exist. However, the strength of the MAX coefficient decreases, suggesting that the sentiment factor drove the standalone MAX effect to a large extent. Further, small-sized and low-priced cryptocurrencies tend to showcase lottery anomaly higher than their counterparts. The availability heuristic is higher in small-cap cryptocurrencies. Due to loss aversion bias, the negative sentiment does not lead to a negative MAX effect. Practical implications It will be useful for the growing investor base in the cryptocurrency market in devising investment strategies. Originality/value The study presents empirical evidence on the impact of behavioral variables on the MAX effect. The study examines the interplay of cryptocurrency investor sentiment and the MAX effect using three novel sentiment proxies.
Decentralized Finance (DeFi) faces critical security challenges due to its pseudonymous and permissionless nature, which exposes it to fraud and market instability. Existing approaches, such as single-agent reinforcement learning (RL) and static graph-based fraud detection, struggle to capture dynamic multi-agent interactions and evolving financial risks. This study proposes an integrated framework combining Multi-Agent Reinforcement Learning (MARL) and Graph Neural Networks (GNNs) to address adaptive decision-making and real-time fraud detection in DeFi. MARL agents, trained in DeepMind’s Melting Pot environment, optimize trading, liquidity provisioning, and arbitrage strategies, while GNNs analyze transaction graphs to detect anomalous patterns. Experimental results demonstrate that MARL agents achieve a 210% increase in average profit per trade and a 57% improvement in market adaptation, alongside a 120% rise in liquidity utilization. The GNN model attains a converged loss below 0.10, reducing false positives by 29%. The integrated system enhances market stability, achieving a stability impact score of 175 within 10 training episodes. This work establishes a scalable, intelligent framework for fraud-resistant trading, cross-chain compliance, and decentralized risk management, advancing the security and efficiency of DeFi ecosystems.
Predicting the price of Bitcoin is crucial, primarily because of the market’s rapid volatility and non-linear environment. For enhanced prediction of the price of Bitcoin, this research proposed a novel interpretable hybrid technique that combines long short-term memory (LSTM) networks with convolutional neural networks (CNN). Deep variational autoencoders (VAE) are used in the stage of preprocessing to determine noticeable patterns in datasets by learning features from historical Bitcoin price data. The CNN-LSTM model additionally implies Shapley additive explanations (SHAP) to promote interpretability and clarify the role of various features. For better performance, the methodology used data cleaning, preprocessing, and effective machine-learning techniques. The hybrid CNN + LSTM model, in collaboration with VAE, obtains a mean squared Error (MSE) of 0.0002, a mean absolute error (MAE) of 0.008, and an R-squared (R2) of 0.99, based on the experimental results. These results show that the proposed model is a good financial forecast method since it effectively reflects the complex dynamics of primary changes in the price of Bitcoin. The combination of deep learning and explainable artificial intelligence improves predictive accuracy as well as transparency, thus qualifying the model as highly useful for investors and analysts.
Purpose: This study aims to examines advanced portfolio management techniques using Long Short-Term Memory (LSTM) networks, the study was applied to investing in cryptocurrencies whose markets are characterized by high-frequency trading, and using behavioral finance models based on the concept of return-risk and deep learning based on the work of artificial neural networks (ANN) and long-term memory (LSTM) algorithms Design/Methodology/Approach: This study adopts quantitative approach. Moreover, A random portfolio consisting of 25 cryptocurrencies was selected based on the database of the website: https://finance.yahoo.com/crypto/ during the period 2021-2024 AD and programming the Python language. And an attempt to evaluate the performance of the models used in accurately predicting the optimal relative weights of the investment portfolio, which proved the relative effectiveness of deep learning models by estimating the values of the mean square error (MSE) at a level of 0.0218% to predict the optimal portfolio weights for 5 days based on training 80% and testing 20% of the study data. Findings: The second hypothesis of this study was accepted, which states the effectiveness of deep learning algorithms to predict the weights of optimal portfolios with a return estimated at 1.7239% and a risk of 1.1219% and a Sharpe index value estimated at 1.5365%, while the Markowitz return-risk model portfolio came with a return rate estimated at 31.15% and a risk of 39.05%. With no diversification of investment on all portfolio assets and a Sharpe index value of 0.7978%. Practical Implications: This study provides important insights that machine learning offers significant advantages in portfolio optimization, from improved forecasting of asset returns to dynamic rebalancing, better risk management, and automation. The ability to handle high-dimensional, non-linear, and non-stationary data makes ML an ideal tool for optimizing portfolios in complex and fast-moving markets; especially in cryptocurrency markets. However, challenges like data quality, overfitting, and interpretability must be addressed to ensure effective deployment of ML in real-world portfolio. Originality/Value: This study provides an original and timely contribution to understanding the use of deep learning for portfolio optimization represents a significant advancement over traditional financial models by offering several original and valuable benefits. These include the ability to capture complex non-linear relationships, dynamic rebalancing in response to real-time data, processing of unstructured data (like sentiment analysis), advanced risk management, and the integration of high-dimensional data. The combination of these capabilities enables more accurate, adaptive, and robust portfolio optimization, ultimately enhancing portfolio performance and reducing risk.
Deep learning and hybrid deep learning models are widely regarded as some of the most effective predictive modeling techniques to date. Their hierarchical architecture enables them to capture complex, non-linear relationships among features and uncover hidden patterns within data, making them particularly powerful for tasks involving high-dimensional and unstructured inputs. But, these models are computationally intensive and require substantial processing time. Moreover, their predictive efficiency is highly dependent on the availability of large-scale datasets. In this study, meta learning model is employed for the prediction of two financial markets: equity market and crypto market. NASDAQ and S&P 500 index has been taken for equity market prediction. On the other hand, Bitcoin & Ethereum are considered for crypto market. Three deep learning models: LSTM, GRU and CNN are trained for the prediction of these four indices and a hybrid deep learning model of GRU and CNN is also developed. Based on RMSE, MAE and R 2 values, it is observed that meta learning yields best results among all trained models with minimum time and using scarce computation resources based on small dataset.
Najma Ali Soomro, Suresh Kumar Oad RAJPUT, Ishfaque Ahmed
Predictions regarding returns and price movements in financial markets can be made using online search engines, which track the sentiments of individual investors. This study aims to analyse how the sentiments of Bitcoin investors impact changes in the American stock market returns. The Bitcoin sentiment index was created to benchmark the sentiments of Bitcoin investors from 2013 to 2018. This index is built by analysing terms from leading business magazines and online journals. Such an index measures potential investors’ sentiments about Bitcoin and how those sentiments impact S&P returns. We use the ordinary least squares method to analyse this. It was found that BSI has a negative impact on S&P returns. Furthermore, the Vector Autoregressive (VAR) model is used to determine the relationship between these economic time series. VAR results indicated a significant positive impact of S&P returns on BSI, while BSI could not predict S&P returns. Consequently, it can be concluded that S&P returns cause changes in BSI. Recognising that Bitcoin sentiment can offer valuable insights and guidance for retail investors during market downturns, much like the S&P 500. By tracking changes in the S&P 500, analysts can anticipate shifts in cryptocurrency market sentiment and take preventative measures when needed. Understanding this relationship is crucial for assessing systemic risks, as volatility in traditional markets can impact the crypto space.
This study aims to develop a hybrid algorithm using the ARIMA model and LSTM-type recurrent neural networks to predict the closing prices of the cryptocurrencies BTC, LTC, and ETH. The methodology includes an exploratory data analysis, followed by the design, implementation, and evaluation of each individual algorithm as well as the combined hybrid algorithm. The results, after experimentation and evaluation of metrics on the test set, indicated that the ARIMA model was inefficient in predicting the closing prices of cryptocurrencies. On the other hand, the hybrid model for BTC showed significant statistical differences in the metrics, with MAE = $726.21 and MAPE = 1.75%, compared to the LSTM model, which achieved MAE = $729.35 and MAPE = 1.76%. These results indicate better performance from the hybrid model. Regarding the RMSE metric, the hybrid model scored 1157.47, while LSTM scored 1159.99; although statistically equivalent, the hybrid model was numerically better. For the remaining metrics and other cryptocurrencies, both methods were statistically equivalent. For five-day-ahead predictions, the hybrid algorithm continued to yield better results for LTC and ETH.
This study introduces a framework that integrates the Hyperledger Fabric blockchain with federated learning to improve cryptocurrency market prediction. Using three major exchanges as distributed nodes, the platform processes trading data and sentiment analysis locally, training machine learning models on each node. The results show that the federated model achieves a prediction deviation of 0. 65% from the actual prices, exceeding the deviation of the centralized LSTM model of 3. 35%. The Hyperledger Fabric network also handles up to 298.7 TPS with zero transaction failures and low latency (0.01s), highlighting the model’s effectiveness for secure and accurate market prediction in the fintech sector.
Financial market efficiency is significantly influenced by the availability and quality of information, with information asymmetry posing a major barrier to optimal market functioning. This article reviews the role of data science in mitigating information asymmetry and enhancing market efficiency, comparing traditional approaches with modern data-driven methods (e.g., machine learning, NLP, and blockchain). It systematically evaluates traditional approaches used to measure and mitigate information asymmetry and highlights their limitations in accurately capturing complex market dynamics. Traditional approaches such as statistical testing, price behavior analysis, and asset pricing models provide fundamental insights but often fail to capture complex, non-linear market dynamics, such as adverse selection, moral hazard, and asset mispricing, due to their reliance on historical data and linear assumptions. In contrast, data science has revolutionized financial market analysis by combining machine learning, natural language processing (NLP), big data analytics, and blockchain technology to solve information imbalances. It enables real-time analysis of unstructured data, improves predictive modeling, and enhances transparency through sentiment analysis, algorithmic trading, and decentralized ledgers. It concludes that integrating data science with traditional finance theory significantly reduces information gaps, offering policymakers and investors tools to foster fairer, more efficient markets. This bridges theoretical finance with computational innovations, demonstrating how data science addresses longstanding limitations in measuring and improving market efficiency.
The reliability and precision of stock market forecasting are of paramount importance to investors, regulatory authorities, and financial institutions.Traditional centralized systems for data processing and model deployment have been found to suffer from critical vulnerabilities, including susceptibility to tampering, single points of failure, and a lack of verifiability.To address these limitations, a novel hybrid framework has been developed that integrates advanced deep learning models with decentralized blockchain infrastructure to ensure both predictive accuracy and data integrity in financial time series forecasting.Temporal dependencies in market dynamics are captured through the use of recurrent neural networks (RNNs) and long short-term memory (LSTM) architectures, which have been extensively trained to model non-linear and non-stationary behaviors in high-frequency financial data.In parallel, a private Ethereum-based blockchain has been deployed to record cryptographic hashes of input datasets, model parameters, and forecasting outputs, thereby ensuring transparency, auditability, and immutability across the data lifecycle.To enable computational scalability, deep learning operations have been executed off-chain, while on-chain mechanisms are utilized for secure checkpointing and traceability.Empirical validation has been conducted using real-time data from the Borsa stanbul (BIST), demonstrating significant improvements in forecasting accuracy when compared with baseline statistical and machine learning (ML) models.Moreover, the integration of blockchain technology has enabled a verifiable audit trail for all predictive operations, enhancing trust in the data pipeline without compromising computational efficiency.The proposed framework represents a significant advancement towards secure, transparent, and trustworthy artificial intelligence (AI) in financial forecasting, with potential implications for the broader decentralized finance (DeFi) ecosystem and regulatory-compliant AI deployments in capital markets.
This study examines and compares the effectiveness of GARCH (Generalized Autoregressive Conditional Heteroskedasticity) and EGARCH (Exponential GARCH) models in forecasting volatility across three distinct financial markets: cryptocurrencies, Indonesian stocks, and U.S. stocks. The research analyzes daily closing price data from April 2018 to September 2024, focusing on five major cryptocurrencies (Bitcoin, Ethereum, Tether, Binance Coin, and Ripple), five Indonesian blue-chip stocks (BBCA, BBRI, BYAN, BMRI, and TPIA), and five major U.S. stocks (Apple, Nvidia, Microsoft, Google, and Amazon). Using comparative analysis of ARCH(1), GARCH(1,1), and EGARCH(1,1,1) models, the study evaluates their predictive accuracy through multiple metrics including AIC, MAE, RMSE, and SMAPE. Results indicate that EGARCH(1,1,1) generally performs better for cryptocurrencies and U.S. stocks, while GARCH(1,1) shows superior performance for Indonesian stocks, suggesting that volatility patterns and optimal forecasting models vary across different market contexts.
Bitcoin is the most traded cryptocurrency by volume and market cap. A number of scholars have directed their research towards characterizing Bitcoin’s speculative behavior using a myriad of techniques such as technical analysis, price regression, and direction classification. For this work, research is conducted using the relatively nascent technique of on-chain data analysis. The goal of this research is to evaluate Bitcoin’s on-chain data in predicting future price direction. First, a classification process of on-chain data features that helps the reader understand their relevance is proposed. To address the curse of dimensionality, feature selection algorithms such as L1 regression, Boruta, and the dimensionality reduction algorithm Principal Component Analysis (PCA) are utilized. The research then explores advanced neural networks for next day price direction prediction, including the Convolutional Neural Network-Long-Short Term Memory (CNN-LSTM) and the Temporal Convolutional Network (TCN). Neural network models and trading strategies are then compared based on their return statistics. A comparative analysis of feature selection, learning model performance, and trading strategy performance is also conducted. Results from the research show that the Boruta feature selection algorithm combined with the CNN-LSTM model performs best compared to other combinations with a prediction accuracy of 82.03% over the testing period. In addition, the on-chain features within the category, realized value, and unrealized value classifications have higher predictive powers for next day price direction prediction. Finally, during trade simulations, the CNN-LSTM model with a Long-Short strategy had an annualized return of 1682.7% and a Sharpe Ratio of 6.47.
Financial institutions increasingly rely on sophisticated database architectures to gain competitive advantages in high-frequency trading and analytics environments. This article examines optimal database technologies for financial applications, comparing in-memory, columnar, time-series, and distributed ledger architectures across standardized financial workloads. Multiple case studies demonstrate how different architectures excel in specific contexts: in-memory processing delivers superior performance for order processing, columnar storage enables faster analytical queries for market analysis, while time-series databases efficiently handle pattern recognition for fraud detection. Performance bottlenecks, consistency trade-offs, regulatory compliance challenges, and security considerations are explored in depth. The results indicate that no single architecture provides optimal performance across all financial application requirements; instead, financial institutions must select technologies based on specific use cases, with heterogeneous architectures often delivering superior results. The article concludes by examining emerging technologies with potential to transform financial database landscapes, including persistent memory, hardware acceleration, specialized indexing structures, AI-integrated engines, and hybrid blockchain solutions.
In this study, we investigate and compare several forecasting models for predicting Bitcoin market prices using historical data sourced from Nasdaq Data Link (formerly Quandl) spanning from 2016 to 2025. Our analysis evaluates traditional time series methods—such as ARIMA and Holt Winters exponential smoothing—alongside modern machine learning and deep learning techniques including LSTM, Prophet, XGBoost, SVR, Random Forest, and GRU. Performance was assessed via metrics such as RMSE, MAE, MAPE, sMAPE, directional accuracy, and R-squared. Our experiments reveal that while classical methods (e.g., ARIMA and Holt Winters) exhibit large estimation errors and limited explanatory capacity, advanced neural network architectures— particularly the GRU—demonstrate superior accuracy with an RMSE of 2,505.84, MAE of 1,760.93, MAPE of 2.79%, and an R-squared of 0.99. The best-performing model (GRU) was deployed as a web application on PythonAnywhere, providing real-time forecasts through an interactive dashboard. This deployment not only validates the predictive efficacy of the GRU model but also offers a practical tool for investors and financial analysts to monitor and predict Bitcoin price movements using reliable Nasdaq data.
The cryptocurrency market is characterized by its high volatility and complex temporal dependencies, posing significant challenges for accurate price prediction. This study introduces advanced hybrid Recurrent Neural Network (RNN) architectures—LSTM-GRU, GRU-BiLSTM, and LSTM-BiLSTM—to enhance the predictive accuracy of cryptocurrency price forecasting. By leveraging the strengths of each RNN variant, the hybrid models effectively capture intricate time-series patterns and nonlinear dependencies inherent in cryptocurrency data. The research follows a comprehensive methodology, including the collection of historical price data for Bitcoin (BTC), Ethereum (ETH), and Litecoin (LTC), rigorous data preprocessing, and the integration of hybrid architectures. Extensive experiments are conducted, and the models are evaluated using key performance metrics, such as Mean Squared Error (MSE), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). Results highlight the superior performance of hybrid RNNs, with LSTM-BiLSTM excelling in BTC price prediction, GRU-BiLSTM and LSTM-GRU demonstrating robust performance for ETH and LTC. This study not only establishes the efficacy of hybrid RNN architectures for time-series forecasting but also underscores their potential for real-world applications in trading strategies. The findings set a new standard for leveraging deep learning in cryptocurrency markets, paving the way for more accurate, reliable, and adaptive forecasting systems. Future work will focus on extending this approach to a broader range of cryptocurrencies and incorporating external market factors to further enhance predictive capabilities.
This study compares the performance of the Temporal Convolutional Network (TCN) and Gated Recurrent Unit (GRU) models in predicting the price of Ethereum, which is important to support cryptocurrency investment strategies. With the high volatility of the cryptocurrency market, an accurate and reliable prediction model is needed. In this study, Ethereum's daily closing price data over four years was analyzed using TCN and GRU models to evaluate its predictive capabilities. Model accuracy is measured using Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Mean Squared Error (MSE). The results showed that the TCN model excelled in average accuracy with lower MAE and MAPE values, while the GRU model showed excellence in reducing the impact of large errors with smaller MSE values. This reflects TCN's superiority in capturing the overall pattern of price movements, while the GRU is more responsive to short-term price fluctuations. These findings demonstrate the potential of both models in cryptocurrency price forecasting, with their respective advantages. This research provides valuable information for investors and researchers in developing predictive strategies in dynamic financial markets. A combination of TCN and GRU models can also be explored to improve prediction performance in the future.