Since late 2019, during one of the largest pandemics in history, COVID-19, global economic recession has continued. Therefore, investors seek an alternative investment that generates profits during this financially risky situation. Cryptocurrency, such as Bitcoin, has become a new currency tool for speculators and investors, and it is expected to be used in future exchanges. Therefore, this paper uses a Value at Risk (VaR) model to measure the risk of investment in Bitcoin. In this paper, we showed the results of the predicted daily loss of investment by using the historical simulation VaR model, the delta-normal VaR model, and the Monte Carlo simulation VaR model with the confidence levels of 99%, 95%, and 90%. This paper displayed backtesting methods to investigate the accuracy of VaR models, which consisted of the Kupiecâs POF and the Kupiecâs TUFF statistical testing results. Finally, Christoffersenâs independence test and Christoffersenâs interval forecasts evaluation showed effectiveness in the predictions for the robustness of VaR models for each confidence level.
S. Ezhilin Freeda, T.C.Ezhil Selvan, I.G. Hemanandhini
Bitcoin is a decentralized digital currency that is completely virtual. It is not managed by a government or a bank. It's a digital file that can be shared from one user to the next. Bitcoin's popularity has risen in recent years, and many people have begun to invest in it. Since investment on bitcoin is increasing day to day and the bitcoin price fluctuates frequently, traders need a way to predict its price in prior so that the risks associated with it can be reduced and capital gain can be improved. The various existing works on price prediction have low accuracy and predict short term price only. Due to the difficulty of determining the exact existence of a Time Series model, it is difficult to generate appropriate forecasts. The Proposed technique uses deep learning to predict bitcoin prices with Recurrent Neural Network model using the time series data to provide the better accuracy. The novelty of work is to obtain a long-term prediction, the recurrent neural network model is trained and tested on the available dataset. This work predicts the value of bitcoin for the year 2021. When compared with other machine learning algorithms like Random Forest, Gaussian NaĂŻve Bayes, Support Vector Machine, K-Nearest Neighbors algorithms the proposed work shows improved accuracy of 76.99% using RNN model.
Bitcoin (cryptocurrencies) is the hottest economic product in recent years. However, due to its highly volatile trend, it is difficult for investors to invest in a targeted manner in the direction of its market trend. And the correlation among cryptocurrencies is often overlooked. In this paper, to solve this problem, the currency data of the past year has been used to put into four regression models to predict and analyze the top five currencies(ranked by market cap) on the market. Among the models, the KNN model has the highest accuracy, reaching 0.923.
Cryptocurrencies are gaining popularity day by day, and their analysis is a fascinating and demanding research topic. The average daily trading volume of Bitcoin was ${\$}67$ billion in May 2021. A peculiar feature of cryptocurrencies is that they are not generally issued by a central authority, making them insusceptible to any governmental impedance. Cryptocurrency rates are closely related to news and influenced by tweets. However, no available dataset can analyze the crypto market adequately. We present CrypTop12, a benchmark dataset for Cryptocurrency Price Movement Prediction based on tweets and historical prices. We collect over 576K tweets related to the top 12 cryptocurrencies, spanning over 1255 days and refine them to filter the tweets that are most relevant to price fluctuations. We also demonstrate use-cases by providing adapted baseline methods and a quantitative results analysis on our dataset.
Olufunke G. Darley, Abayomi Isiaka O. Yussuff, Adetokunbo A. Adenowo
Abstract This paper investigated Bitcoin daily closing price using time series approach to predict future values for financial managers and investors. Daily data were sourced from CoinDesk, with Bitcoin Price Index (BPI) for 5 years (January 1, 2016 to May 31, 2021) extracted. Data analysis and modelling of price trend using Autoregressive Integrated Moving Average (ARIMA) model was carried out, and a suitable model for forecasting was proposed. Results showed that ARIMA(6,1,12) model was the most suitable based on a combination of number of significant coefficients and values of volatility, Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). A two-month test window was used for forecasting and prediction. Results showed a decline in prediction accuracy as number of days of the test period increased; from 99.94% for the first 7 days, to 99.59 % for 14 days and 95.84% for 30 days. For the two-month test period, percentage accuracy was 84.75%. The study confirms that the ARIMA model is a veritable planning tool for financial managers, investors and other stakeholders; especially for short-term forecasting. It is however imperative that the influence of external factors, such as investorsâ/influencersâ comments and government intervention, that may affect forecasting be taken into consideration.
Abstract Cryptocurrencies are digital assets that can be stored and transferred electronically. Bitcoin (BTC) is one of the most popular cryptocurrencies that has attracted many attentions. The BTC price is considered as a high volatility time series with non-stationary and non-linear behavior. Therefore, the BTC price forecasting is a new, challenging, and open problem. In this research, we aim the predicting price using machine learning and statistical techniques. We deploy several robust approaches such as the Box-Jenkins, Autoregression (AR), Moving Average (MA), ARIMA, Autocorrelation Function (ACF), Partial Autocorrelation Function (PACF), and Grid Search algorithms to predict BTC price. To evaluate the performance of the proposed model, Forecast Error (FE), Mean Forecast Error (MFE), Mean Absolute Error (MAE), Mean Squared Error (MSE), as well as Root Mean Squared Error (RMSE), are considered in our study.
Recently, Bitcoin has gained great importance in the cryptocurrency market with the highest market capitalization. Investors and researchers have attempted to find out the drivers of Bitcoin prices and if they are predictable. However, there is only limited research in the literature that identifies the most effective economic and technical variables for predicting Bitcoin prices using machine learning models. Thus, in this study, the future Bitcoin prices utilizing several economic and technical factors using the ANFIS model are aimed to forecasted between 01.05.2013 - 26.02.2021 periods. The findings show that the ANFIS model produced accurate and consistent predicting results that are in line with the real data. As a result, investors who wish to make a profit by predicting future Bitcoin values might consider using the ANFIS approach as a forecasting tool.
Aleksandar PetroviÄ, Ivana Strumberger, Timea Bezdan, Hothefa Shaker Jassim ¡ 5 authors
Cryptocurrencies are defined as digital mediums of exchange, that use strong cryptography for securing the transactions and verifying the ownership of the coins. Blockchain operates in the background to guarantee the security, transparency and traceability of the transactions. Consequently, cryptocurrencies became more and more popular and established their considerable presence in financial sector. However, one of the major drawbacks in the cryptocurrency market is the unreliability and unpredictability of their values, that poses a major risk for any kind of investment. Predicting the price of cryptocurrencies is therefore a hot research domain today. This paper proposes a novel method to predict the prices, that is based on a hybrid machine learning and swarm intelligence approach. The results of the conducted experiments suggest that the proposed model obtains higher accuracy than other recent similar approaches, and that it can be successfully applied for this important task.
We study the information dynamics between the largest Bitcoin exchange markets during the bubble in 2017-2018. By analysing high-frequency market-microstructure observables with different information theoretic measures for dynamical systems, we find temporal changes in information sharing across markets. In particular, we study the time-varying components of predictability, memory, and synchronous coupling, measured by transfer entropy, active information storage, and multi-information. By comparing these empirical findings with several models we argue that some results could relate to intra-market and inter-market regime shifts, and changes in direction of information flow between different market observables.
This research is the first attempt to customize a trading system that is based on second order stochastic dominance (SSD) to five known cryptocurrenciesâ daily data: Bitcoin, Ethereum, XRP, Binance Coin, and Cardano. Results show that our system can predict price trends of cryptocurrencies, trade them profitably, and in most cases outperform the buy and hold (B&H) simple strategy. Our systemâs best performance was achieved trading XRP, Binance Coin, Ethereum, and Bitcoin. Although our system has also generated a positive net profit (NP) for Cardano, it failed to outperform the B&H strategy. For all currencies, the system better predicted long trends than short trends.
This paper initially presents a brief overview of the cryptocurrency and its history. We discuss the novel nature of literature attempting to create hybrid artificial neural network models to predict prices of cryptocurrency. For the remaining majority of the paper, we present the details of various hybrid artificial neural networks that have successfully been implemented to predict cryptocurrency prices in the form of a survey. Comparison of methods and results follow in the results section.
In this paper, I examine how social media affects cryptocurrencies and more traditional stocks. I use data on Twitter posts in combination with daily stock prices to estimate the causal effect of a tweet on stock and coin prices. To do this, I use a difference-indifference regression with index funds as my control group, which allows me to capture general market trends that coins and stocks would follow if not for intervention. I find that tweets have a significant impact on cryptocurrencies that last up to three days after the post. The increase in coin prices is driven by tweets from Tyler Winklevoss and tweets about Tezos and Ethereum specifically. Meanwhile, Twitter posts have no impact on more traditional stocks. These results suggest that social media can provide the public with valuable information in real time for fast moving and volatile crypto assets, while their effects on more stable and institutionalized traditional stocks are more muted.
Atif Naseer, Enrique Nava Baro, Sultan Daud Khan, Y. Vila ¡ 5 authors
In recent years, cryptocurrency has become gradually more significant in economic regions worldwide. In cryptocurrencies, records are stored using a cryptographic algorithm. The main aim of this research was to develop an optimal solution for predicting the price of cryptocurrencies based on user opinions from social media. Twitter is used as a marketing tool for cryptoanalysis owing to the unrestricted conversations on cryptocurrencies that take place on social media channels. Therefore, this work focuses on extracting Tweets and gathering data from different sources to classify them into positive, negative, and neutral categories, and further examining the correlations between cryptocurrency movements and Tweet sentiments. This paper proposes an optimized method using a deep learning algorithm and convolution neural network for cryptocurrency prediction; this method is used to predict the prices of four cryptocurrencies, namely, Litecoin, Monero, Bitcoin, and Ethereum. The results of analyses demonstrate that the proposed method forecasts prices with a high accuracy of about 98.75%. The method is validated by comparison with existing methods using visualization tools.
M. Kabir Hassan, Fahmi Ali Hudaefi, Rezzy Eko Caraka
Purpose This paper aims to explore netizenâs opinions on cryptocurrency under the lens of emotion theory and lexicon sentiments analysis via machine learning. Design/methodology/approach An automated Web-scrapping via RStudio is performed to collect the data of 15,000 tweets on cryptocurrency. Sentiment lexicon analysis is done via machine learning to evaluate the emotion score of the sample. The types of emotion tested are anger, anticipation, disgust, fear, joy, sadness, surprise, trust and the two primary sentiments, i.e. negative and positive. Findings The supervised machine learning discovers a total score of 53,077 sentiments from the sampled 15,000 tweets. This score is from the artificial intelligence evaluation of eight emotions, i.e. anger (2%), anticipation (18%), disgust (1%), fear (3%), joy (15%), sadness (3%), surprise (7%), trust (15%) and the two sentiments, i.e. negative (4%) and positive (33%). The result indicates that the sample primarily contains positive sentiments. This finding is theoretically significant to measure the emotion theory on the sampled tweets that can best explain the social implications of the cryptocurrency phenomenon. Research limitations/implications This work is limited to evaluate the sampled tweetsâ sentiment scores to explain the social implication of cryptocurrency. Practical implications The finding is necessary to explain the recent phenomenon of cryptocurrency. The positive sentiment may describe the increase in investment in the decentralised finance market. Meanwhile, the anticipation emotion may illustrate the publicâs reaction to the bubble prices of cryptocurrencies. Social implications Previous studies find that the social signals, e.g. word-of-mouth, netizensâ opinions, among others, affect the cryptocurrenciesâ movement prices. This paper helps explain the social implications of such dynamic of pricing via sentiment analysis. Originality/value This study contributes to theoretically explain the implications of the cryptocurrency phenomenon under the emotion theory. Specifically, this study shows how supervised machine learning can measure the emotion theory from data tweets to explain the implications of cryptocurrencies.
The machine learning method has been used in stock price prediction for a long time, and the price of cryptocurrencies such as bitcoin has attracted more and more attention in recent years. This paper aims to improve the method applicable to the stock market and try to use it in cryptocurrency price prediction. A simple three-layered feedforward artificial neural networks (ANN) model was applied in this paper to predict the daily directions of cryptocurrency prices. The historical trading data of Bitcoin, Ethereum, and Cardano were used in the experiments. Nine selected technical indicators were preprocessed into discrete trend data, and they were input into the model together with three additional indicators for training. This study has preliminarily obtained an effective result with price prediction accuracy of the three cryptocurrencies between 61% and 65%.
Purpose The purpose of this study is to compare five data-driven-based ML techniques to predict the time series data of Bitcoin returns, namely, alternating model tree, random forest (RF), multiple linear regression, multi-layer perceptron regression and M5 Tree algorithms. Design/methodology/approach The data used to forecast time series data of Bitcoin returns ranges from 8 July 2010 to 30 Aug 2020. This study used several predictors to predict bitcoin returns including economic policy uncertainty, equity market volatility index, S&P returns, USD/EURO exchange rates, oil and gold prices, volatilities and returns. Five statistical indexes, namely, correlation coefficient, mean absolute error, root mean square error, relative absolute error and root relative squared error are determined. The results of these metrices are used to develop colour intensity ranking. Findings Among the machine learning (ML) techniques used in this study, RF models has shown superior predictive ability for estimating the Bitcoin returns. Originality/value This study is first of its kind to use and compare ML models in the prediction of Bitcoins. More studies can be carried out by using further cryptocurrencies and other ML data-driven models in future.
Evan Millikan, Preethi Subramanian, Minnu Helen Joseph
Cryptocurrencies are non-physical currency that solely exist as represented by 0s and 1s within the world of computers. One of the most popular cryptocurrencies in the market right now being Bitcoin, was first invented to solve the inherent problem with using traditional currency when purchasing online. However, unexpectedly Bitcoin soon found itself to be one of the most profitable investment opportunities to be hedged on with its yearly growth unrivaled by any traditional investment product such as stocks, bonds, or real-estate. However, unlike the stock market which has been the subject of multitude of research papers, the cryptocurrency market has not been treated the same way and as such there is still a huge opportunity open in this industry. Thus, Bitcoin Vision wants to utilize this opportunity and propose the use of machine learning and deep learning architecture to predict the price movement trend of bitcoin (up or down) for the short-term prediction and predict the price of bitcoin for the long-term prediction. Being able to predict the future of the market prove to be useful in the stock market and as such this paper decide to replicate that opportunity to be presented in the cryptocurrency market as well. In this literature review paper, we have proposed the comparison of state-of-the-art deep learning model such as Long Short-Term Memory (LSTM) with traditionally successful machine learning model such as Random Forest and ARIMA to find out which model provide the best result.
Aboosaleh Mohammad Sharifi, Kaveh KhaliliâDamghani, Farshid Abdi, Soheila Sardar
Cryptocurrencies are considered as new financial and economic tools having special and innovative features, among which Bitcoin is the most popular. The contribution of the Bitcoin market continues to grow due to the special nature of Bitcoin. The investors' attention to Bitcoin has increased significantly in recent years due to significant growth in its prices. It is important to create a prediction system which works well for investment management and business strategies due to the high chaos and volatility of Bitcoin prices. In this study, in order to improve predictive accuracy, Bitcoin price dataset is first divided into a time interval through time window, then propose a new model based on Long Short-Term Memory (LSTM) neural networks and Metaheuristic algorithms. Chaotic Dolphin Swarm Optimization algorithm is used to optimize the LSTM. Performance evaluation indicated that the proposed model can have more effective predictions and improve prediction accuracy. In addition, the performance of the optimized model is better and more reliable than other models.
Noura Metawa, Mohamemd I. Alghamdi, Ibrahim M. ElâHasnony, Mohamed Elhoseny
Recently, bitcoin-based blockchain technologies have received significant interest among investors. They have concentrated on the prediction of return and risk rates of the financial product. So, an automated tool to predict the return rate of bitcoin is needed for financial products. The recently designed machine learning and deep learning models pave the way for the return rate prediction process. In this aspect, this study develops an intelligent return rate predictive approach using deep learning for blockchain financial products (RRP-DLBFP). The proposed RRP-DLBFP technique involves designing a long short-term memory (LSTM) model for the predictive analysis of return rate. In addition, Adam optimizer is applied to optimally adjust the LSTM modelâs hyperparameters, consequently increasing the predictive performance. The learning rate of the LSTM model is adjusted using the oppositional glowworm swarm optimization (OGSO) algorithm. The design of the OGSO algorithm to optimize the LSTM hyperparameters for bitcoin return rate prediction shows the novelty of the work. To ensure the supreme performance of the RRP-DLBFP technique, the Ethereum (ETH) return rate is chosen as the target, and the simulation results are investigated in different measures. The simulation outcomes highlighted the supremacy of the RRP-DLBFP technique over the current state of art techniques in terms of diverse evaluation parameters. For the MSE, the proposed RRP-DLBFP has 0.0435 and 0.0655 compared to an average of 0.6139 and 0.723 for compared methods in training and testing, respectively.
Bitcoin is firmly becoming a mainstream asset in our global society. Its highly volatile nature has traders and speculators flooding into the market to take advantage of its significant price swings in the hope of making money. This work brings an algorithmic trading approach to the Bitcoin market to exploit the variability in its price on a day-to-day basis through the classification of its direction. Building on previous work, in this paper, we utilise both features internal to the Bitcoin network and external features to inform the prediction of various machine learning models. As an empirical test of our models, we evaluate them using a real-world trading strategy on completely unseen data collected throughout the first quarter of 2021. Using only a binary predictor, at the end of our three-month trading period, our models showed an average profit of 86\%, matching the results of the more traditional buy-and-hold strategy. However, after incorporating a risk tolerance score into our trading strategy by utilising the model's prediction confidence scores, our models were 12.5\% more profitable than the simple buy-and-hold strategy. These results indicate the credible potential that machine learning models have in extracting profit from the Bitcoin market and act as a front-runner for further research into real-world Bitcoin trading.