S. Venkatesh, B Rashmitha, S Manjunadha, Md Junaid
A type of digital currency known as a cryptocurrency allows all transactions to be completed online. There is no hard cash version of this soft currency. We highlight that a decentralized currency differs from a centralized currency in the any user of a virtual currency can purchase services without the need for third parties to get involved. Due to its extreme price volatility, using these cryptocurrencies has an impact on trade and international relations. Moreover, the constantly fluctuating oscillations indicate the urgent need for a more precise method of predicting this price. Deep learning techniques that use effective learning models for training data, including the LSTM, GRU, and Feedback Neural Network, can be used to do this. Benchmark datasets are used to test the suggested strategy. That brings us to the neural network, one of the clever data mining technologies that researchers in many domains have been using for the past ten years. In the current economy, stock market data is essential. There are two types of forecasting methodologies: nonlinear models (ARCH, GARCH, Neural Network) and linear models (AR, MA, ARIMA, ARMA). To forecast a company's stock price based on past prices, we employed the Box Jenkins Model also known as ARIMA, and Long Short-Term Memory (LSTM), and Feedback Neural Network also known as RNN.
This research paper aims to conduct a time series forecasting of the bitcoin mean weighted price using the data from Kaggle. The data has a one-minute resolution and includes the following variables: timestamp, open, high, low, close, volume (BTC), volume (currency), and weighted price. Data analysis was achieved using R, a statistical computing and graphics programming language. The main findings of this research paper were that the bitcoin mean weighted price had a strong upward trend and exhibited high volatility over time. The time series also had weak seasonal and significant random components, indicating periodic fluctuations and noise in the data. Four years of data were used to estimate the mean change for the following month. The results indicate that while the expected value may rise somewhat, it will do so with significant variability and unpredictability. The main implications of this research paper were that there was a potential for profit or loss depending on the timing and strategy of buying or selling bitcoins.
People are starting to see the cryptocurrency market as a viable source of income and investment, similar to the stock market, as the concept of cryptocurrencies continues to gain popularity. Predicting Bitcoin returns is related to financial machine learning, which uses time series to forecast price variance. This study starts with the daily close price of Bitcoin for its initial dataset. The price is transformed into percentages and binary classes, which categorize into “Up” and “Down”, after which a time series is applied to produce two datasets: a categorical dataset for classification and a numerical dataset for regression. For classification that represents a Binary classification in asset-price forecasting, k-fold cross-validation is applied to ensure that the best classifiers are selected for testing and analysis. Most of the regression analysis was based on visualization, which displayed the predicted prices by each regressor in front of the original values and helped analyse the models’ results more accurately. The outcomes of this study were achieved by anticipating bitcoin returns using classification and regression machine learning models, despite the approaches’ low accuracy and significant precision rate to the “Up” class. At this stage, with a significant limitation regarding the dataset and a lack of other indicators, a model capable of predicting future variations is considered a beneficial addition for many trading tools or even for crypto market analysts.
Anomaly detection is critical in the financial sector, especially as financial environments evolve with increasing digitization, posing challenges for real -time anomaly detection. Recently, deep learning (DL) algorithms have emerged as promising solutions for this problem. This study presents a DL -based anomaly detection model utilizing various algorithms, including LSTM, GRU, and 1dCNN, applied to Tesla's stock market and Ethereum cryptocurrency data sets. Hyperparameter optimization is performed using grid search. Results show that the GRU algorithm achieves the highest prediction score in both datasets, while the 1dCNN algorithm performs the lowest. Additionally, anomaly values are graphically demonstrated using GRU for both datasets. Accurate bookkeeping is essential for legitimate business operations, yet the complexity of financial auditing requires new solutions. Supervised and unsupervised machine learning techniques are increasingly applied to detect fraud and anomalies in accounting data. This paper addresses the challenge of detecting financial misstatements in general ledger (GL) data, proposing seven supervised ML techniques, including deep learning, and two unsupervised ML techniques. Models are trained and evaluated on real -life GL datasets, demonstrating high potential in detecting predefined anomaly types and efficiently sampling data. Practical implications of these solutions in accounting and auditing contexts are discussed. The rapid development of computer networks brings both convenience and security challenges due to various abnormal flows. Traditional detection systems, like intrusion detection systems (IDS), have limitations, necessitating real -time updates to function effectively. With the advent of machine learning and data mining, new methods for abnormal network flow detection have emerged. This paper introduces the random forest algorithm for detecting abnormal samples, proposing the concept of an abnormal point scale to measure sample abnormality based on similarity. Simulation experiments demonstrate the superiority of random forest -based detection in terms of model accuracy and computing efficiency compared to other methods.
Abbas M. Ali, Abdullah Shah, Azaz Hassan Khan, Malik Umar Sharif · 8 authors
Software-based technical indicators have been widely used for the stock market forecasting, aiming to predict market direction. Even though many algorithms for the software based technical indicators are presented, there are almost no hardware implementations reported in the literature. In this paper, the hardware implementation is presented for three commonly used technical indicators: Moving Average Convergence/Divergence (MACD), Relative Strength Index (RSI), and Aroon. Latency evaluation is conducted for Bitcoin and Ethereum within a single-day timeframe, utilizing the Xilinx Zynq-7000 programmable SoC XC7Z020-CLG484-1 platform. Additionally, various hardware/software (HW/SW) partitioning strategies are explored to leverage the flexibility of software alongside the performance advantages of hardware via the Zynq SoC platform. The results show that the best performing technical indicator is MACD with a speedup of 30 times over its software only counterpart. Furthermore, a hybrid design integrating multiple technical indicators is proposed, pairing MACD with RSI due to their competitive throughput values, differing by only 0.38 microseconds. This hybrid approach capitalizes on the parallel processing capabilities of hardware, enabling multiple systems to operate simultaneously.
Cryptocurrencies have revolutionized the financial landscape by providing decentralized and anonymous payment systems, making them an intriguing subject for investors and researchers. This article delves into applying machine learning techniques for predicting cryptocurrency prices, mainly focusing on Bitcoin, Ethereum, and Binance Coin. Employing a range of machine learning models, including XGBoost, Linear Regression, and Gaussian Processes, the study aims to evaluate their predictive performance comprehensively. The results are promising; our models outperform existing studies, achieving impressively low RMSE values of 0.0040 for Bitcoin, 0.028 for Ethereum, and 0.027 for Binance Coin. These findings contribute valuable insights into the volatility and dynamics of cryptocurrency prices and underscore the potential of machine learning in shaping financial decision-making. Future directions include integrating advanced deep learning models, additional data sources, and ensemble methods to enhance prediction accuracy and robustness.
The decentralised nature of cryptocurrency, coupled with its potential for significant financial returns, has elevated its status as a sought-after investment opportunity on a global scale. Nonetheless, the inherent unpredictability and volatility of the cryptocurrency market present considerable challenges for investors aiming to forecast price movements and secure profitable investments. In response to this challenge, the current investigation was conducted to assess the efficacy of three Machine Learning (ML) algorithms, namely, Gradient Boosting (GB), Random Forest (RF), and Bagging, in predicting the daily closing prices of six major cryptocurrencies, namely, Binance, Bitcoin, Ethereum, Solana, USD, and XRP. The study utilised historical price data spanning from January 1, 2015 to January 26, 2024 for Bitcoin, from January 1, 2018 to January 26, 2024 for Ethereum and XRP, from January 1, 2021 to January 26, 2024 for Solana, and from January 1, 2019 to January 26, 2024 for USD. A novel approach was adopted wherein the lagging prices of the cryptocurrencies were employed as features for prediction, as opposed to the conventional method of using opening, high, and low prices, which are not predictive in nature. The data set was divided into a training set (80%) and a testing set (20%) for the evaluation of the algorithms. The performance of these ML algorithms was systematically compared using a suite of metrics, including R2, adjusted R2, Mean Square Error (MSE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE). The findings revealed that the GB algorithm exhibited superior performance in predicting the prices of Bitcoin and Solana, whereas the RF algorithm demonstrated greater efficacy for Ethereum, USD, and XRP. This comparative analysis underscores the relative advantages of RF over GB and Bagging algorithms in the context of cryptocurrency price prediction. The outcomes of this study not only contribute to the existing body of knowledge on the application of ML algorithms in financial markets but also provide actionable insights for investors navigating the volatile cryptocurrency market.
Mohammad Vahidpour, Amir Daneshvar, Mohsen Amini Khouzani, Mahdi HOMAYOUNFAR
Financial markets are characterised by their dynamic, non-linear, and fluctuating nature. Analysing financial time series in these contexts is a complex and challenging task. Candlestick patterns are recognised as among the most widely used financial tools and offer invaluable insights into market sentiment and psychology. However, manual analysis of these patterns presents significant challenges. Therefore, leveraging machine learning methods becomes a necessity for overcoming these challenges. In this study, a four-step framework was introduced in which the data preparation process is executed on the price data of the 20 cryptocurrencies. Forty-eight candlestick patterns were extracted alongside returns. Employing the long shortterm memory (LSTM) neural network, structured with multiple layers, each specialising in a specific cryptocurrency, enables individualised prediction of market returns. Evaluation of model accuracy and sensitivity is conducted via the confusion matrix, and two distinct trading strategies assess the capital portfolio. The research findings underscore the profitability of the proposed model across all scenarios. Candlestick patterns serve as powerful tools for understanding market sentiments and identifying shifts in market trends. However, their standalone efficacy is limited. Integrating them with other technical analysis tools facilitates more informed decision-making and fosters a deeper understanding of market dynamics.
Son yıllarda, para piyasalarında ve bankacılık sektöründe yaşanan krizlerin etkisiyle merkezi para otoritelerine olan güven sarsılmış ve bu nedenle merkezi olmayan bir sistem arayışına girilmiştir. Bu vesile ile, 1998 yılında ilk kripto para kavramı gündeme gelmiştir. Kripto piyasaları, kripto para birimlerinin alınıp satıldığı dijital veya sanal pazar yerlerini ifade etmektedir. Kripto para birimleri, güvenli finansal işlemler için kriptografi kullanan merkezi olmayan dijital varlıklar olarak karşımıza çıkmaktadır. 2009 yılına gelindiğinde ise ana kripto para birimi olan Bitcoin ile ilk işlem gerçekleştirilmiştir. Merkezi olmayan bu sanal paranın zaman içerisinde talebinde gözlemlenen artış ile birlikte piyasa değeri de hızla yükselmiştir. Çalışmanın amacı, piyasa değeri her geçen gün artan bu kripto para piyasalarının yapısını açıklamak ve ana kripto para birimi olan Bitcoin özelinde fiyat hareketlerinin seyrini incelemektir. 2015:02 - 2022:01 dönemleri arasında aylık veri setinin kullanıldığı analizde GPH yöntemi kullanılmıştır. Elde edilen sonuçlara göre, Bitcoin serisinin dirençli bir yapı sergilediği gözlenmektedir. Bitcoin fiyatlarını etkileyen herhangi bir iktisadi şokun belirli bir süre etkisini sürdüreceği ancak uzun dönemde etkisini kaybedeceği sonucuna varılmıştır.
The research presented in this paper is the first to introduce a thorough Descriptive-Predictive–Prescriptive (DPP) Framework for comprehending the interaction between social media and cryptocurrencies. Recognizing the underexplored domain of the social-media–cryptocurrency interaction, we delve into its many aspects, better understanding present dynamics, forecasting potential future trajectories, and prescribing best solutions for stakeholders. We evaluate social media speech and behavior connected to cryptocurrencies using big data analytics, translating raw data into meaningful insights using Natural Language Processing (NLP) techniques like sentiment analysis. When applied to an experimental dataset, the DPP nets superior results compared to the baseline approach, displaying an improvement of 3.44% of the Root Mean Square Error (RMSE) metric and 4.59% of the Mean Absolute Error (MAE) metric. The unique DPP framework enables a more in-depth assessment of social media’s influence on cryptocurrency trends, and lays the path for strategic decision-making in this nascent but rapidly developing field of study.
As a matter of fact, stock market prediction remains a challenging and crucial aspect of investment decision-making. Contemporarily, cryptocurrencies are one of the hottest underlying assets on account of its high volatility. In this case, based on high accuracy prediction, it is available to achieve large extra return from the crypto markets. With this in mind, this study investigates the usage of machine learning algorithms, including LSTM, GRU, as well as bi-LSTM, for predicting cryptocurrency prices, focusing on Bitcoin (BTC), Ethereum (ETH), as well as Litecoin (LTC). According to the analysis, the study reveals that the GRU model consistently outperforms other algorithms, MAPE and RMSE values across all three cryptocurrencies. These findings underscore the reliability and efficiency of the GRU model in cryptocurrency price prediction. Furthermore, the research compares the model's performance with previous studies, reaffirming its effectiveness and potential for practical application in investment strategies as well as decision-making.
With the growing concern of AI safety, there is a need to trust the computations done by machine learning (ML) models. Blockchain technology, known for recording data and running computations transparently and in a tamper-proof manner, can offer this trust. One significant challenge in deploying ML Classifiers on-chain is that while ML models are typically written in Python using an ML library such as Pytorch, smart contracts deployed on EVM-compatible blockchains are written in Solidity. We introduce Machine Learning to Smart Contract (ML2SC), a PyTorch to Solidity translator that can automatically translate multi-layer perceptron (MLP) models written in Pytorch to Solidity smart contract versions. ML2SC uses a fixed-point math library to approximate floating-point computation. After deploying the generated smart contract, we can train our models off-chain using PyTorch and then further transfer the acquired weights and biases to the smart contract using a function call. Finally, the model inference can also be done with a function call providing the input. We mathematically model the gas costs associated with deploying, updating model parameters, and running inference on these models on-chain, showing that the gas costs increase linearly in various parameters associated with an MLP. We present empirical results matching our modeling. We also evaluate the classification accuracy showing that the outputs obtained by our transparent on-chain implementation are identical to the original off-chain implementation with Pytorch.
Joy Dip Das, Ruppa K. Thulasiram, A. Thavaneswaran
This work addresses the intricate task of predicting the prices of diverse financial assets, including stocks, indices, and cryptocurrencies, each exhibiting distinct characteristics and behaviors under varied market conditions. To tackle the challenge effectively, a novel hybridized architecture, AE-GRU, integrating the encoder-decoder principle with GRU is designed. The experimentation involves multiple activation functions and hyperparameter tuning. With extensive experimentation and enhancements applied to AE-LSTM, the proposed AE-GRU architecture still demonstrates significant superiority in forecasting the annual prices of volatile financial assets from multiple sectors mentioned above. Thus, the novel AE-GRU architecture emerges as a superior choice for price prediction across diverse sectors and fluctuating market scenarios by extracting important non-linear features of financial data and retaining the long-term context from past observations.
Kripto para birimleri, 2009 yılında ortaya çıkmalarından bu yana oldukça popüler hale gelmiştir. Özellikle Bitcoin'in 3 Ocak 2009'da piyasaya sürülmesinden sonra, diğer kripto para birimlerinin piyasaya çıkışı hız kazanmıştır. Bu popülerlik artışının ardından, kripto para birimlerinin tahmini önemli bir konu haline gelmiştir. Bu çalışmanın ana amacı, Bitcoin (BTC), Ethereum (ETH) ve Binance (BNB) kripto para getirilerini öngörmek için geleneksel zaman serisi yöntemlerinden olan ARIMA-GARCH ile birlikte LSTM (Long Short-Term Memory) derin öğrenme yaklaşımını kullanarak elde edilen tahmin performanslarını karşılaştırmaktır. Bu çerçevede, çalışma literatüre yeni bir katkı sunmayı amaçlamaktadır. Her bir kripto para birimi için farklı zaman aralıklarında günlük veriler kullanılmış ve bu veriler %90 eğitim ve %10 test verisi olarak bölünmüştür. Çalışmada, yöntemler RMSE ve MSE değerlendirme kriterleri kullanılarak karşılaştırılmıştır. Genel olarak, BTC serisinde ARIMA-GARCH yöntemi eğitim verisinde daha iyi sonuçlar gösterirken, test verisi için LSTM yöntemi daha etkili olmuştur. BNB serisinde ise hem eğitim hem de test verisi için LSTM yöntemi daha üstün performans sergilemiştir. ETH serisinde ise her iki veri seti için ARIMA-GARCH yöntemi daha iyi sonuçlar ortaya koymaktadır. Bu çalışma, finansal veri tahmininde her iki yöntemin de önemli bir performans sergileyebildiğini vurgulamaktadır.
Gabriele Morello, Mojtaba Eshghie, Sofia Bobadilla, Martin Monperrus
The DISL dataset features a collection of $514,506$ unique Solidity files that have been deployed to Ethereum mainnet. It caters to the need for a large and diverse dataset of real-world smart contracts. DISL serves as a resource for developing machine learning systems and for benchmarking software engineering tools designed for smart contracts. By aggregating every verified smart contract from Etherscan up to January 15, 2024, DISL surpasses existing datasets in size and recency.
Bitcoin (BTC) represents an emerging asset class, offering investors an alternative avenue for diversification across various units of exchange. The recent global banking crisis of 9 March 2023 has provided an opportunity to reflect on how Bitcoin’s perception as a speculative asset may be evolving. This paper analyzes the volatility behavior of BTC in comparison to gold and the traditional financial market using GARCH models. Additionally, we have developed and incorporated a bank index within our volatility analysis framework, aiming to isolate the impact of financial crises while minimizing idiosyncratic risk. The aim of this work is to understand Bitcoin’s perception among investors and, more importantly, to determine whether BTC can be considered a new asset class. Our findings show that in terms of volatility and price, BTC and gold have responded in very similar ways. Counterintuitively, the financial market seems not to have experienced high volatility and significant price swings in response to the March 9th crisis. This suggests a consumer tendency to seek refuge in both Bitcoin and gold.
Cryptocurrency, a digital currency managed by decentralized networks, has gained immense popularity since the inception of Bitcoin. These digital assets, often characterized by extreme price volatility, have generated substantial interest from investors. Traditional financial models struggle to account for the unique dynamics and complexities of cryptocurrencies, prompting the adoption of deep learning techniques. This study investigates the use of Long Short-Term Memory (LSTM), Neural Networks, and Deep Learning (CNN) in predicting cryptocurrency prices. These deep learning models leverage various data sources, such as technical indicators and sentiment analysis, to gain a comprehensive understanding of cryptocurrency markets. The research evaluates the performance of these models using Root Mean Squared Error (RMSE) as the primary metric. The results demonstrate that the hybrid model, combining LSTM, Neural Networks, and Deep Learning, exhibits the highest predictive accuracy across multiple cryptocurrencies, including Bitcoin (BTC), Ethereum (ETH), and Binance Coin (BNB). However, challenges persist, such as model adaptability to unforeseen market events and data noise. Future developments may involve incorporating external factors and interdisciplinary collaboration to create more holistic valuation models. Despite these challenges, the study underscores the potential of hybrid deep learning models in enhancing cryptocurrency valuation accuracy and their relevance in risk management strategies for investors and traders.
The valuation and prediction of cryptocurrency prices have become increasingly important in the financial market. Therefore, this study aims to focus on the selection and evaluation of machine learning models for cryptocurrency valuation. Thus, two types of machine learning models, gradient boosting trees (Xgboost and LightGBM) and neural networks, are compared to determine their effectiveness in generating features for cryptocurrency valuation. Additionally, correlation tests are conducted to identify the most suitable input variables for the models. The results demonstrate that the generated features have a significant impact on the accuracy of machine learning predictions for cryptocurrency prices. It highlights the potential of machine learning models in accurately predicting and evaluating the value of cryptocurrencies. Overall, the findings of this study contribute to the understanding of the role of machine learning in cryptocurrency valuation and provide valuable insights for investors and researchers. By leveraging machine learning techniques, investors can make informed decisions and develop effective investment strategies in the cryptocurrency market. This study contributes to cryptocurrency valuation research. Leveraging machine learning enables informed decisions and effective investment strategies. Furthermore, the findings inform the development of advanced machine learning models and algorithms for cryptocurrency valuation.
Yuxin Zhang, Rajiv Garg, Linda L. Golden, Patrick L. Brockett · 5 authors
Cryptocurrencies like Bitcoin have received substantial attention from financial exchanges. Unfortunately, arbitrage-based financial market price prediction models are ineffective for cryptocurrencies. In this paper, we utilize standard machine learning models and publicly available transaction data in blocks to predict the direction of Bitcoin price movement. We illustrate our methodology using data we merged from the Bitcoin blockchain and various online sources. This gave us the Bitcoin transaction history (block IDs, block timestamps, transaction IDs, senders’ addresses, receivers’ addresses, transaction amounts), as well as the market exchange price, for the period from 13 September 2011 to 5 May 2017. We show that segmenting publicly available transactions based on investor typology helps achieve higher prediction accuracy compared to the existing Bitcoin price movement prediction models in the literature. This transaction segmentation highlights the role of investor types in impacting financial markets. Managerially, the segmentation of financial transactions helps us understand the role of financial and cryptocurrency market participants in asset price movements. These findings provide further implications for risk management, financial regulation, and investment strategies in this new era of digital currencies.
In light of recent cryptocurrency value fluctuations, Bitcoin is gradually gaining recognition as an investment vehicle. Given the market's inherent volatility, accurate forecasting becomes crucial for making informed investment decisions. Notably, previous research has utilized machine learning methods to enhance the accuracy of Bitcoin price predictions. However, few studies have explored the potential of employing diverse modeling methods for sampling with varying data formats and dimensional characteristics. This study aims to identify the internal feature subset that yields the highest returns in forecasting Bitcoin's price. Specifically, Bitcoin's internal features were categorized into four groups: currency data, block details, mining information, and network difficulty. Subsequently, a long short-term memory (LSTM) artificial neural network was employed to predict the next day's Bitcoin closing price, utilizing various categorizations of feature subsets. The model underwent training using two and a half years of historical data for each feature. The findings revealed a mean absolute error rate of 6.38% when modeling with the block details category features. This enhanced performance primarily stemmed from the positive relationship between Bitcoin price and this data subset's low ambiguity. Experimental results underscored that, compared to other investigated feature subsets, the categorization of block detail features provided the most accurate Bitcoin price predictions, laying the foundation for future research in this domain.
High Accessibility and Easy Investment makes Cryptocurrency an important income source for many people. Cryptocurrency is a kind of Digital/Virtual currency which is created using blockchain Technology and is protected by Cryptography. Cryptocurrencies enables users to Accept, Transfer and request the capital between the Users without the requirement of intermediaries such as banks. Now a day many Cryptocurrencies are available across the world such as Bitcoin, Litecoin, Monero, Dogecoin etc. This study is more determined over a very famous and demanding Cryptocurrency known as Bitcoin over the past years. Here, firstly we make an effort to predict the price of bitcoin by examining numerous numbers of parameters that affect the cost of bitcoin. Different kinds of Machine learning models will be used to estimate the price of Bitcoin. This study provides the accuracy and precision of each model that are used in this study and determine the suitable method to estimate the price more accurately.
In the dynamic world of finance, the application of Artificial Intelligence (AI) in pair trading strategies is gaining significant interest among scholars. Current AI research largely concentrates on regression analyses of prices or spreads between paired assets for formulating trading strategies. However, AI models typically exhibit less precision in regression tasks compared to classification tasks, presenting a challenge in refining the accuracy of pair trading strategies. In pursuit of high-performance labels to elevate the precision of classification models, this study advanced the Triple Barrier Labeling Method for enhanced compatibility with pair trading strategies. This refinement enables the creation of diverse label sets, each tailored to distinct barrier configurations. Focusing on achieving maximal profit or minimizing the Maximum Drawdown (MDD), Genetic Algorithms (GAs) were employed for the optimization of these labels. After optimization, the labels were classified into two distinct types: High Risk and High Profit (HRHP) and Low Risk and Low Profit (LRLP). These labels then serve as the foundation for training machine learning models, which are designed to predict future trading activities in the cryptocurrency market. Our approach, employing cryptocurrency price data from 9 November 2017 to 31 August 2022 for training and 1 September 2022 to 1 December 2023 for testing, demonstrates a substantial improvement over traditional pair trading strategies. In particular, models trained with HRHP signals realized a 51.42% surge in profitability, while those trained with LRLP signals significantly mitigated risk, marked by a 73.24% reduction in the MDD. This innovative method marks a significant advancement in cryptocurrency pair trading strategies, offering traders a powerful and refined tool for optimizing their trading decisions.
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.
Since 2012, researchers have explored various factors influencing Bitcoin prices. Up until the end of July 2023, more than 9100 research papers on cryptocurrencies were published and indexed in the Web of Science Clarivate platform. The objective of this paper is to analyze the impact of publications on Bitcoin prices. This study aims to uncover significant themes within these research articles, focusing on cryptocurrencies in general and Bitcoin specifically. The research employs latent Dirichlet allocation to identify key topics from the unstructured abstracts. To determine the optimal number of topics, perplexity and topic coherence metrics are calculated. Additionally, the abstracts are processed using BERT-transformers and Word2Vec and their potential to predict Bitcoin prices is assessed. Based on the results, while the research helps in understanding cryptocurrencies, the potential of academic publications to influence Bitcoin prices is not significant, demonstrating a weak connection. In other words, the movements of Bitcoin prices are not influenced by the scientific writing in this specific field. The primary topics emerging from the analysis are the blockchain, market dynamics, transactions, pricing trends, network security, and the mining process. These findings suggest that future research should pay closer attention to issues like the energy demands and environmental impacts of mining, anti-money laundering measures, and behavioral aspects related to cryptocurrencies.