This study presents a novel methodology for multi-step Bitcoin (BTC) price prediction by combining advanced stacking-based architectures with temporal attention mechanisms. The proposed Temporal Attention-Enhanced Stacking Network (TAESN) integrates the complementary strengths of diverse machine learning algorithms while emphasizing critical temporal features, leading to substantial improvements in forecasting accuracy over traditional methods. Comprehensive experimentation and robust evaluation validate the superior performance of TAESN across various BTC prediction horizons. Additionally, the model not only demonstrates enhanced predictive accuracy but also offers interpretable insights into the temporal dynamics underlying cryptocurrency markets, contributing to both practical forecasting applications and theoretical understanding of market behavior.
Stock price prediction is a challenging research topic because of non-linearity, significant noise and volatility of time series data.Deep learning techniques enable to learn complex and non-linear patterns of sequential time series data.Long Short-Term Memory (LSTM) is a technique which is designed to handle time series data.While LSTM model is used to extract temporal dependencies of stock data, the performance can be limited by noisy data and the challenge of capturing intricate patterns.In this research, LSTM-based framework with residual unit and attention mechanism is proposed to enhance the temporal dependencies and important features of stock price movements.Residual unit with skip connection captures more complex patterns and representations in stock price data and reduces the over-fitting problem to noisy time series data.LSTM with attention focuses on the significant time stamps which enhances the model prediction performance.The proposed system is experimented on five datasets: Apple (AAPL), Bitcoin, Ethereum, Litecoin and GOLD_PRICE.To prove the effectiveness of the model, the proposed system is compared with LSTM and Bidirectional LSTM (Bi-LSTM) models.Experimental results show that the proposed system outperforms baseline models such as LSTM, Bi-LSTM, LSTM+Bi-LSTM and state-of-the-art methods in term of error rates such as mean square error, root mean square error and mean absolute error.
Forecasting the price of cryptocurrencies is a notoriously hard and significant problem, due to the rapid market growth and high volatility. In this article, we propose a methodology for predicting future values of cryptocurrency exchange rates by developing a Non-linear Autoregressive with Exogenous Inputs (NARX) prediction model that uses the most adequate external information. The exogenous variables considered are historical values of the exchange rate and a series of technical indicators. The selection of the most relevant external inputs is based on the computation of the mutual information indicator and estimated using the k-nearest neighbor method. The methodology employs a fine-tuned Long Short-Term Memory (LSTM) neural network as the regressor. We have used quantitative and trend accuracy measures to compare the proposed method against other state-of-the-art LSTM-based models. In addition, regarding the input selection process, the proposed approach was compared against the most commonly used one, which is based on the cross-correlation coefficient. A long series of experiments and statistical analyses proved that the proposed methodology is highly accurate and the resulting model outperforms the state-of-the-art LSTM-based models.
Zhuohuan Hu, F. Richard Yu, Zizhou Zhang, Haoran Zheng · 6 authors
This paper leverages machine learning algorithms to forecast and analyze financial time series. The process begins with a denoising autoencoder to filter out random noise fluctuations from the main contract price data. Then, one-dimensional convolution reduces the dimensionality of the filtered data and extracts key information. The filtered and dimensionality-reduced price data is fed into a GANs network, and its output serve as input of a fully connected network. Through cross-validation, a model is trained to capture features that precede large price fluctuations. The model predicts the likelihood and direction of significant price changes in real-time price sequences, placing trades at moments of high prediction accuracy. Empirical results demonstrate that using autoencoders and convolution to filter and denoise financial data, combined with GANs, achieves a certain level of predictive performance, validating the capabilities of machine learning algorithms to discover underlying patterns in financial sequences. Keywords - CNN;GANs; Cryptocurrency; Prediction.
Ponzi schemes deceive investors with promises of high returns, relying on funds from new investors to pay earlier ones, creating a misleading appearance of profitability. These schemes are inherently unsustainable, collapsing when new investments wane, leading to significant financial losses. Many researchers have focused on detecting such schemes, but challenges remain due to their evolving nature. This study proposes a novel hybrid machine-learning approach to enhance Ponzi scheme detection. Initially, we train an XGBoost classifier and extract its features. Meanwhile, we tokenize opcode sequences, train a gated recurrent unit (GRU) model on these sequences, and extract features from the GRU. By concatenating the features from the XGBoost classifier and the GRU, we train a final XGBoost model on this combined feature set. Our methodology, leveraging advanced feature engineering and hybrid modeling, achieves a detection accuracy of 96.57%. This approach demonstrates the efficacy of combining XGBoost and GRU models, along with sophisticated feature engineering, in identifying fraudulent activities in Ethereum smart contracts. The results highlight the potential of this hybrid model to offer more robust and accurate Ponzi scheme detection, addressing the limitations of previous methods.
We examine whether disagreement in social media discussions related to financial markets affects subsequent volatility and abnormal trading volume. We also compare how traditional and digital asset markets differ by comparing stocks and Bitcoin. We show that social media disagreement is positively associated with future market volatility and abnormal trading volume in the stock market. The effect of disagreement is more pronounced at the individual stock level than at the index level. A higher level of social media disagreement also increases the probability of extremely negative stock market returns. In contrast, disagreement in Bitcoin-related social media weakly affects subsequent volatility but does not affect trading volume or extremely negative returns. Our findings also reveal that market activity impacts the disagreement in the stock market and Bitcoin communities differently.
Dünyada en çok rağbet gören kripto para birimi olması nedeniyle Bitcoin (BTC), yatırımcılar ve araştırmacılar için son yıllarda dikkat çekici hale gelmiştir. Merkezi bir para birimi olmaması ve spekülasyonlara açık olması BTC fiyatında yüksek oynaklığa sebep olmaktadır. BTC fiyatının oynaklığının dikkate alınarak tahminlenmesi özellikle yatırımcılar için büyük önem taşımaktadır. Son yıllarda Makine Öğrenmesi (ML) yöntemlerinin gelişmesiyle birlikte birçok finansal alanda olduğu gibi kripto paraların fiyat tahminlemesinde sıklıkla ML yöntemlerine başvurulmaktadır. ML yöntemleri geleneksel ekonometrik yöntemlerin aksine veri setinde meydana gelen dalgalanmaları herhangi bir varsayıma ihtiyaç duymadan dikkate almakta ve çoğu zaman daha iyi sonuçlar vermektedirler. Bu çalışmada, 01.01.2018 ile 21.12.2023 tarihleri arasında BTC fiyatı geleneksel ekonometrik yöntem olan ARIMA ile ML yöntemi olan LSTM kullanılarak tahminlenmeye çalışılmıştır. Yöntemler karşılaştırılırken performans kriterleri olarak RMSE, MAE ve MAPE kriterleri kullanılmıştır. Çalışmanın sonuçlarına göre LSTM yöntemi en düşük RMSE ve MAPE değerlerine sahip olmuştur.
Cryptocurrencies fluctuate in markets with high price volatility, which becomes a great challenge for investors. To aid investors in making informed decisions, systems predicting cryptocurrency market movements have been developed, commonly framed as feature-driven regression problems that focus solely on historical patterns favored by domain experts. However, these methods overlook three critical factors that significantly influence the cryptocurrency market dynamics: 1) the macro investing environment, reflected in major cryptocurrency fluctuations, which can affect investors’ collaborative behaviors, 2) overall market sentiment, heavily influenced by news, which impacts investors’ strategies, and 3) technical indicators, which offer insights into overbought or oversold conditions, momentum, and market trends are often ignored despite their relevance in shaping short-term price movements. In this paper, we propose a dual prediction mechanism that enables the model to forecast the next day’s closing price by incorporating macroeconomic fluctuations, technical indicators, and individual cryptocurrency price changes. Furthermore, we introduce a novel refinement mechanism that enhances the prediction through market sentiment-based rescaling and fusion. In experiments, the proposed model achieves state-of-the-art performance (SOTA), consistently outperforming ten comparison methods in most cases. Our code and data can be found at https://github.com/aamitssharma07/SAL-Cryptopulse
Tatiana Silveira Camacho, Guilherme Jonas da Silva
As a highly speculative asset, bitcoin’s (BTC) demand is particularly based on the perceptions of agents about the financial asset. Especially institutions that began leaving a bigger footprint in the market, causing changes in transaction flow and price cycles. To assess how halving dynamics changed the BTC market, Wavelet methodology was carried out with daily data (from January 2011 to December 2021), on price and transaction count. Decomposition in scale and frequency indicate that flows were altered by the arrival of new investors, and stronger correlations between prices and transactions were found at lower frequencies (longer time horizon).
Igba Emmanuel, Moral Kuve Ihimoyan, Babatunde Awotiwon, Akinkunmi Rasheed Apampa
This paper explores the integration of advanced machine learning models, including BERT, GPT, and the Prophet algorithm, with finance investment strategies to enhance predictive modeling and trend analysis in blockchain technology. The rapid evolution of blockchain has transformed financial ecosystems, offering decentralized platforms for secure and transparent transactions. However, predicting market trends and investment opportunities within this domain remains a complex challenge due to high volatility and the multifaceted nature of financial data. By leveraging the natural language processing capabilities of BERT and GPT for sentiment analysis and market behavior prediction, combined with the time-series forecasting strength of the Prophet algorithm, this study aims to provide a robust framework for analyzing blockchain-driven financial markets. Furthermore, the integration of finance investment strategies ensures practical applicability by aligning machine learning insights with real-world investment decision-making processes. The proposed approach demonstrates potential for optimizing portfolio management, enhancing risk mitigation, and improving strategic investment in blockchain ecosystems. This work bridges the gap between cutting-edge machine learning technologies and financial innovation, offering valuable insights for researchers and practitioners in both domains.
Md Zahidul Islam, Md. Shahidul Islam, Md Abdullah Al Montaser, Md Rasel · 7 authors
The cryptocurrency market is one of the most dynamic and volatile markets in the world's financial ecosystem, and investment landscapes in the US financial market have changed so much. In slightly over a decade, cryptocurrencies have moved from niche digital assets to mainstream investment opportunities such as Bitcoin, Ethereum, and many others. The prime objective of this research project was to investigate the effectiveness of various machine learning algorithms in the prediction of cryptocurrency prices within the volatile US financial market. This research pinpointed which Machine Learning techniques provide the most accurate and reliable predictions under different market conditions, with a full understanding of their strengths and limitations. The dataset gathered for analyzing and forecasting cryptocurrency prices entailed diverse and extensive data points, affirming a well-rounded foundation for machine learning algorithms. Particularly, current and historic price data from cryptocurrency exchanges such as Binance, Coinbase, and Kraken, together with trading metrics important for the definition of market dynamics. Aggregated data from financial databases such as Coin-Market-Cap, Crypto-Compare, and Yahoo Finance comes in structured form and presents historical consistency, hence perfectly fitting for machine learning applications. Models considered for the study ranged from simple, linear methods to complex ensemble and gradient-boosting algorithms. Precise performance evaluation is a proxy of its reliability and correctness of effectiveness in price predictions in a cryptocurrency market. Several measures of the effectiveness of prediction have been used here for assessing the different properties of models' performance: Precision, Recall, and F1-Score. Additional performance metrics were applied to evaluate the models in this study including Mean Absolute Error, Root Mean Squared Error, and R-squared. The gradient Boosting model did an excellent job as compared to other algorithms, as the values of accuracy, precision, recall, and F1-score for both classes were quite high. All three models have quite a relatively low MAE and RMSE, which means that each model is remarkably good at predicting the target variable. The application of machine learning models in the sphere of cryptocurrency price prediction might finally give very important implications to investors and stakeholders of the financial market in the USA, especially since recently, cryptocurrencies have been made integral parts of both individual and institutional investors' portfolios and trading strategies. To investors, it may provide indications of the entry and exit points, diversification of portfolios, and risk management by using machine learning models. Consolidation with the financial system will indeed mark a strategic shift toward data-driven decision-making in investment management and trading by integrating machine learning models into the financial systems.
Sibtain Syed, Syed Muhammad Umar Talha, Arshad Iqbal, Naveed Ahmad · 5 authors
Cryptocurrency is recognized as a leading digital currency by its peer-to-peer transfer capabilities and secure features. Accurately forecasting cryptocurrency price trends holds substantial significance for investors and traders, as they inform critical decisions regarding the acquisition, divestment, or retention of cryptocurrencies, guided by expectations of value, risk assessment, and potential returns. This study also aims to identify a resourceful technique to efficiently forecast prices of cryptocurrencies such as Bitcoin (BTC), Binance (BNB), Ripple (XRP), and Tether (USDT) using optimal data-driven models (LSTM, GRU, and BiLSTM models) using bias correction. The proposed methodology includes collecting cryptocurrency data and precious metal data from Coindesk and BullionVault, respectively, and then finding the optimal model input combination for each cryptocurrency by lag adjustment and correlating feature selection. Hyperparameter tuning was performed by trial-and-error technique, and an early stopping function was applied to minimize time and space complexity. Bias correction (BC) is applied to model-forecasted price trends to reduce errors in evaluation and to enhance accuracy by adjusting model outputs to reduce prediction bias, providing a refined alternative to traditional unadjusted deep learning methods. GRU-BC outperformed other models in forecasting Bitcoin (with MAE 25.291, RMSE 31.266, MAPE 2.999) and USDT (with MAE 0.0006, RMSE 0.0012, MAPE 0.0622) price trends, while BiLSTM-BC was superior in predicting XRP (with MAE 0.0129, RMSE 0.0171, MAPE 2.9013) and BNB (with MAE 2.2759, RMSE 2.8357, MAPE 1.9785) market price flow.
The global crypto market has garnered attention from investors due to its high return potential, yet the extreme volatility and associated risks pose significant concerns. This study aims to analyze the risks and returns of cryptocurrency investments using a quantitative approach on global crypto market data. The methods employed include volatility analysis using the GARCH model, as well as risk and return evaluation through CAPM. The results indicate that cryptocurrencies exhibit significantly higher volatility compared to traditional assets, with substantial return potential but accompanied by high risk. These findings imply the need for stricter diversification strategies and risk management for crypto investors to optimize returns and minimize losses.
This study examines the bitcoin price in USD in the world by developing a suitable time series model to identify its future trends. This data set consists of monthly bitcoin prices from August 2010 to July 2024. It was found that the original series is not stationary and not seasonality. The stationary was achieved by the first difference. Of the parsimonious models identified based on the Partial Autocorrelation Function (PACF) and Autocorrelation Function (ACF) of the stationary series, an auto-regressive integrated moving average (ARIMA) (2,1,2) model was identified as the best-fitt ed model. The significance of the model and its parameters and information criteria such as the Akaike Information Criterion (AIC), Schwarz Criterion, and log-likelihood was used to identify the best-fitted model. The model was trained using data from August 2010 to March 2024. The residuals of the model were found to be white noise. The mean absolute percentage error (MAPE) for validation data is 7.09%. The percentage errors for the validating set are all positive and varied from 3.5% to 12.9%. The predicted Bitcoin price (USD) from August to October 2024 are $59947.88, $60308.7, and $60669.53. Bitcoin price can be utilized by market demand and supply, regulatory environment, and technology development. Keywords: ACF; ARIMA models; Bitcoin price; Forecasting; PACF; Time series analysis
Muhammad Yousaf, Muhammad Imran Tariq, Abdul Jabbar, Syed Qaiser Jalil
This comprehensive review paper explores the diverse landscape of cryptocurrency forecasting, tracing its evolution from an alternative to traditional monetary systems to its significant growth in the global financial arena. It consolidates existing research by categorizing and analyzing 234 scholarly articles, organizing them into machine learning, deep learning, deep reinforcement learning, and statistical methodologies, and evaluating the related metrics. The case study titled “Examining the performance differences between backtesting and forward testing” highlights the challenges investors face, as strategies that appear effective in backtesting often fail in practical use. Another case study, “Social Data Exploration in Cryptocurrency Trends,” examines how social media data can provide insights into market movements and investor sentiment, revealing the impact of social trends on cryptocurrency prices. The findings section provides a detailed view, illuminating trends such as yearly publication rates, methodological distributions, input features, training/testing splits, the total number of data samples considered, and forecasting time horizons. This survey paper serves as a valuable resource, providing researchers and investors with a solid foundation for understanding and navigating the dynamic field of cryptocurrency forecasting.
The aim of this article is to examine the reasons why cryptocurrency volatility hinders its potential to replace fiat money as legal tender. We focus on Bitcoin and Ethereum for this analysis. By applying an augmented Dickey-Fuller stationarity test, we demonstrate that cryptocurrencies lack a long-term trend; instead, their movement is erratic and highly volatile. Furthermore, eGARCH models indicate that volatility tends to decrease and is expected to persist in this pattern. In summary, theoretical and empirical analysis suggests that, due to their nature based solely on supply and demand and their high volatility, cryptocurrencies are not suitable as primary investment instruments or stores of value.
Li Yi Thong, Ricky Chee Jiun Chia, Mohd Fahmi Ghazali
Research Question: Does uncertainty indices have impact on cryptocurrency? Motivation: Most of the previous study investigate the impact of geopolitical risk and economic policy uncertainty on Bitcoin only and less research investigate the long run and short run relationship between the uncertainty indices and cryptocurrency. Hence, this study investigates whether the economic policy uncertainty, geopolitical risk and US equity market uncertainty have an impact on Bitcoin, Ethereum and Binance Coin by the multivariate VAR Granger non-causality. Idea: This study applied three different uncertainty indices (geopolitical risk, economic policy uncertainty and US equity market uncertainty) and top three ranking cryptocurrency (Bitcoin, Ethereum and Binance Coin) to investigate and compare the impact of uncertainty indices on cryptocurrency with different uncertainty conditions and applied top three ranking cryptocurrency in cryptocurrency market to reinforce the result. Data: This study applied monthly data with 42 observations which cover the period of December 2017 until May 2021 and data for cryptocurrency extracted from investing.com, while the uncertainty indices from policyuncertainty.com. Method/Tools: This study utilize multivariate VAR Granger non-causality to examine the cointegration relationship between the cryptocurrency and uncertainty indices. Findings: The results show that the economic policy uncertainty, geopolitical risk and US equity market uncertainty cointegrated with Bitcoin, while Binance Coin cointegrated with geopolitical risk only. Hence, the economic policy uncertainty, geopolitical risk and US equity market uncertainty plays a vital role in the Bitcoin prediction and geopolitical risk plays an important role to forecast the Binance Coin. Contributions: The Bitcoin investors may focus on the changes in economic policy uncertainty, geopolitical risk and US equity market uncertainty to predict the Bitcoin return, and Binance Coin investors focus on the geopolitical risk.
Cryptocurrencies are known for their volatility and instability, making them an attractive but risky investment for traders, analysts, and researchers. As the allure of Bitcoin (BTC) and other cryptocurrencies continues to grow, so does the interest in predicting their prices. To forecast the market rate and sustainability of cryptocurrencies, this study uses machine learning-based time series analysis. The study employs forecast periods ranging from 1 to 10-minutes to categorize the consistency of the market. High-frequency pricing of cryptocurrencies is anticipated with a timestep of up to 10 seconds using various deep learning (DL) models. A hybrid model combining long short-term memory (LSTM) and gated recurrent unit (GRU) is created and compared with standard LSTM and GRU models. Mean squared error (MSE) is the benchmark for estimating the models' performance. The study achieves better results than benchmark models, with MSE values for BTC, Cardano (ADA), and Cosmos (ATOM) in a 5-minute window size being 0.000192, 0.000414, and 0.000451, respectively, and for a 10-minute window size being 0.000212, 0.000197, and 0.000746. Compared to existing models, the suggested model offers a high price predicting accuracy. This study on crypto price prediction using machine learning applications is a preliminary investigation into the topic.
Moiz Qureshi, Hasnain Iftikhar, Paulo Canas Rodrigues, Mohd Ziaur Rehman · 5 authors
Bitcoin (BTC-USD) is a virtual currency that has grown in popularity after its inception in 2008. BTC-USD is an internet communication network that makes using digital money, including digital payments, easy. It offers decentralized clearing of transactions and money supply. This study attempts to accurately anticipate the BTC-USD prices (Close) using data from September 2023 to September 2024, comprising 390 observations. Four machine learning models—Multi-layer Perceptron, Extreme Learning Machine, Neural Network AutoRegression, and Extreme-Gradient Boost—as well as four time series models—Auto-Regressive Integrated Moving Average, Auto-Regressive, Non-Parametric Auto-Regressive, and Simple Exponential Smoothing models—are used to achieve this end. Various hybrid models are then proposed utilizing these models, which are based on simple averaging of these models. The data-splitting technique, commonly used in comparative analysis, splits the data into training and testing data sets. Through comparison testing with training data sets consisting of 30%, 20%, and 10%, the present work demonstrated that the suggested hybrid model outperforms the individual approaches in terms of error metrics, such as the MAE, RMSE, MAPE, SMAPE, and direction accuracy, such as correlation and the MDA of BTC. Furthermore, the DM test is utilized in this study to measure the differences in model performance, and a graphical evaluation of the models is also provided. The practical implication of this study is that financial analysts have a tool (the proposed model) that can yield insightful information about potential investments.
V. SimhadriAppanna, M. Manohara, B.V. Sai Thrinath, D. Leela Rani · 6 authors
The proliferation of mobile devices and personal computing has revolutionized stock and crypto currency trading. While many struggle with navigating trading intricacies, adept practitioners find lucrative opportunities for wealth accumulation. Automated price prediction systems, particularly the Long Short-term Memory (LSTM) model, offer passive trading approaches, eliminating exhaustive decision-making processes. Acquiring and organizing data, followed by rigorous calculations and analysis, culminates in accurate price forecasts. Though not infallible, these models discern trends and project crypto currency trajectories. Notably, Bitcoin serves as a prime example. These systems offer invaluable insights, aiding investors in strategic decision-making amid the dynamic crypto currency landscape.
Investment in cryptocurrencies has garnered substantial attention in the recent past as the prices for these digital currencies started recording all-time highs. While there are numerous contenders in the cryptocurrency market, bitcoin has emerged to be the most popular and sought after digital currency. Despite its popularity, the theoretical understanding of the value of this cryptocurrency is still limited. Hence this study aims to find out the significant predictors of the bitcoin price and build a machine-learning based model to evaluate and predict the complex phenomenon of bitcoin price. Here we contribute to the extant literature by searching for the potential contributors of bitcoin prices ranging from fundamental, macroeconomic, financial, speculative, and technical sources to the most marked event of 2020 i.e., Covid19 pandemic. For this purpose, we have used state-of-the-art machine learning, deep learning, and statistical time-series models (univariate and multivariate) to forecast bitcoin price. The study revealed that deep learning models performed almost at par with Random Forest model for both pre- and whilst-Covid19 era. Traditional time-series models, namely VAR and VECM gave the most consistent performance within acceptable margins for both pre- and whilst-Covid era. We have also found that macroeconomic factors play an important role in determining bitcoin price formulation process during both periods, while mining difficulty and market sentiment factors gain more importance during pre-Covid period. In addition, number of covid cases is also found to be a significant factor for the prediction of bitcoin price during whilst-Covid period.
The behavior of the Bitcoin market is dynamic and erratic, impacted by a range of elements including news developments and investor mood. One well-known aspect of bitcoin is its extreme volatility. This study uses both conventional econometric techniques and deep learning algorithms to anticipate the volatility of Bitcoin returns. The research is based on historical Bitcoin price data spanning October 2014 to February 2022, which was obtained using the Yahoo Finance API. In this work, we contrast the efficacy of generalized autoregressive conditional heteroskedasticity (GARCH) and threshold ARCH (TARCH) models with long short-term memory (LSTM), bidirectional LSTM (Bi-LSTM), and multivariate Bi-LSTM models. Model effectiveness is evaluated by means of root mean squared error (RMSE) and root mean squared percentage error (RMSPE) scores. The multivariate Bi-LSTM model emerges as mostly effective, achieving an RMSE score of 0.0425 and an RMSPE score of 0.1106. This comparative scrutiny contributes to understanding the dynamics of Bitcoin volatility prediction, offering insights that can inform investment strategies and risk management practices in this quickly changing environment of finance.
This study presents a comparative analysis of two advanced attention-based deep learning models—Attention-LSTM and Attention-GRU—for predicting Bitcoin price movements. The significance of this research lies in integrating moving average technical indicators with deep learning models to enhance sensitivity to market momentum, and in normalizing these indicators to accurately reflect market trends and reversals. Utilizing historical OHLCV data along with four key technical indicators (SMA, EMA, TEMA, and MACD), the models classify trends into uptrend, downtrend, and neutral categories. Experimental results demonstrate that the inclusion of technical indicators, particularly MACD, significantly improves prediction accuracy. Furthermore, the Attention-GRU model offers computational efficiency suitable for real-time applications, while the Attention-LSTM model excels in capturing long-term dependencies. These findings contribute valuable insights for financial forecasting, providing practical tools for cryptocurrency traders and investors.