In today’s era of big data, deep learning and artificial intelligence have formed the backbone for cryptocurrency portfolio optimization. Researchers have investigated various state of the art machine learning models to predict Bitcoin price and volatility. Machine learning models like recurrent neural network (RNN) and long short-term memory (LSTM) have been shown to perform better than traditional time series models in cryptocurrency price prediction. However, very few studies have applied sequence models with robust feature engineering to predict future pricing. In this study, we investigate a framework with a set of advanced machine learning forecasting methods with a fixed set of exogenous and endogenous factors to predict daily Bitcoin prices. We study and compare different approaches using the root mean squared error (RMSE). Experimental results show that the gated recurring unit (GRU) model with recurrent dropout performs better than popular existing models. We also show that simple trading strategies, when implemented with our proposed GRU model and with proper learning, can lead to financial gain.
Bitcoin, merkezi bir otoriteye veya finansal bir kuruluşa bağlı olmayan ve kriptografik özellikler içeren dijital (kripto) paralardan biridir. Bitcoin’ in Merkezi otoriteye bağlı olmaması ve fiyatını etkileyen faktörlerin arz ve talep ile açıklanması yüksek volatite ile sonuçlanmıştır. Son dönemlerde yatırımcıların en büyük endişesi fiyatlardaki aşırı volatilite durumudur. Çalışmada Blockchain Teknolojisi, Madencilik ve Blockchain Teknolojisinin bir çıktısı olan Bitcoin kısaca anlatılmıştır. Çalışmanın uygulama bölümünde literatürde sıklıkla kullanılan yöntemlerden olan ve asimetrik volatilitenin belirlenmesi amacıyla ARCH, GARCH, ARCHM, EGARCH ve TARCH modelleri kullanılmıştır. Bu amaçla Bitcoin/USD kuru kapanış fiyatlarından Bitcoine ilişkin tarihsel getiriler hesaplanmıştır. Hesaplama dönemi 01.01.2015-11.02.2018 olarak belirlenmiştir. Yapılan analizler sonucunda volatilite tahmini için en iyi sonuç veren TARCH yöntemi bulunmuştur.
The main purpose of this study in determine whether Bitcoin is becoming an alternative investment option compared to other financial instruments in Turkey. To perform analysis for this purpose, we created sample includes Bitcoin, Bist 100 National Index, Bond, Euro and Gold. We calculated daily returns in terms of % for each type of instrument for the period range from 02.02.2012 to 17.12.2018. Methodology -In the study, we tested the stationarity of the series and the causality relationships between the series through the nonlinear unit root and causality tests. Stationarity of the series and causality relations between them are determined with the help of charts. Both in unit root test and in causality test, firstly we obtained test statistics and normalized them by using critical values then transferred them to the charts. In order to decide about when test statistics is higher than critical values null hypothesis is rejected. Findings-Bitcoin, Usd, Gold and Bond are found non-stationary for the whole period. We find out that returns of Euro is more stable than Usd. Bitcoin has been fluctuating during all period except for first months of 2013 while Gold's returns are intensively volatile when politic and economic risks are increasing. This show investors in Turkey still consider Gold as safe port for their investments. Usd currency is volatile for all period because of its increasing demand all over the world. In causality test we observed that there is a causality relation from Bitcoin's returns to Bist 100 returns. This result shows that Bitcoin is becoming an alternative (substitute) investment tool for domestic and foreign investors compared to Borsa Istanbul. There is more causality relation between Bitcoin and Gold than Bitcoin and Bist 100. Conclusion-Based on the findings of this analysis, , it may be accepted that Bitcoin has been becoming an alternative investment/savings tool for the Turkey case. Regulatory institutions should create a required legal framework for the Bitcoin because Bitcoin has a great potential to prevent of tax evasion, the terminate the informal economy and eliminate intermediation costs.
Dijital para birimleri son yıllarda etkinliğini artırarak uluslararası piyasalarda önemli oranda talep görmeye başlamıştır. Dijital para birimleri arasında işlem hacmi ve getiri oranları dikkate alındığında, Bitcoin ön plana çıkmaktadır. Çalışmada, Bitcoin fiyatları ile döviz kurları arasındaki ilişki incelenmektedir. Bu doğrultuda kurulan model kapsamında, 24 Kasım 2013 - 04 Mart 2018 dönemlerini kapsayan haftalık veriler ile yapısal kırılmalı testler kullanılarak döviz kurları ile Bitcoin fiyatları arasındaki ilişki irdelenmiştir. Analizlerde kullanılan verilerin I (1) düzeyinde durağan olduğu tespit edilmiş olup yapısal kırılmaya izin veren Maki Eşbütünleşme testi sonuçlarına göre değişkenler arasında yapısal kırılmalarla birlikte uzun dönemli bir ilişki olduğu sonucuna ulaşılmıştır. Hacker-Hatemi-J Bootstrap Nedensellik testi sonuçlarında ise dolar kurundan Bitcoin fiyatlarına doğru %1 anlamlılık düzeyinde nedensellik ilişkisi tespit edilmiştir.
Roberto B. Corcino, Karl Patrick Casas, Allan Roy Elnar
In this paper, a mathematical model is constructed that would capture the pattern of the MSD of fluctuations of Bitcoin unit prices over time in daily basis by applying the method of White Noise Analysis. The raw data of 2805 ordered points (t,p) are used in the study, which are taken from coindesk.com, where p is the Bitcoin unit price at given time t.
Kripto para olarak da adlandırılan dijital para fiyatlarındaki değişimler son yıllarda yatırımcıların oldukça ilgisini çekmiştir. Hızlı fiyat değişimlerinden getiri elde etmek isteyen yatırımcılar yeni bir varlık olan dijital paralara yönelmişlerdir. Bu doğrultuda, dijital paraların geleneksel menkul kıymetlerine alternatif olma ihtimalleri tartışılmaya başlanmıştır. Çalışmada, Bitcoin fiyatları ile Borsa İstanbul arasındaki eşbütünleşme ve nedensellik ilişkisini tespit etmek amaçlanmıştır. Bu kapsamda, Engle-Granger ve Gregory-Hansen eşbütünleşme testleri ile Toda-Yamamoto ve Hacker-Hatemi-J nedensellik testlerinden faydalanılmıştır. Bulgular, her iki eşbütünleşme testine göre Bitcoin fiyatları ile Borsa İstanbul endeks değeri arasında orta ve uzun vadede bir eşbütünleşme ilişkisinin olmadığını; nedensellik testlerinden sadece Toda-Yamamoto nedensellik testine göre Borsa İstanbul’dan Bitcoin fiyatlarına doğru tek yönlü nedensellik ilişkisi olduğunu göstermiştir.
Cryptocurrencies have gained tremendous popularity over the past few years. The purpose of this study is to try to understand the factors that are driving cryptocurrency-related trading activities. Focusing on the well-established cryptocurrency called Bitcoin, we find that online search popularity and the volume of trade in unrelated stock markets positively and negatively, respectively, influence Bitcoin trading volume. We also find no statistical evidence that the underlying sentiment behind relevant financial news influence Bitcoin trading volume. We believe these results might be of great value to investors interested in cryptocurrencies and might instigate further research on this topic.
Purpose - This study conducts an analysis to reveal the interaction between Bitcoin and Exchange Rates to find out whether Bitcoin is becoming a substitution for the exchange rates.Methodology - To investigate the mutually interaction between the exchange rates and the Bitcoin, the interaction (relationship) between daily closing price of both exchange rates and Bitcoin was analyzed through the Var model. Thus, it was tried to show the sensitivity of the values of Bitcoin to the changes occured in the exchange rates.Findings - Based on Variance Decomposition analysis, BITCOIN and Euro can be considered as largely external variables and their prices are not significantly affected by USD. An interesting result in this study is that the USD exchange rate was found to be significantly sensitive to the Euro.Conclusion - Findings obtained from analysis show that Bitcoin and Excange Rates have not become an alternative tools for each other yet.
The value of various Cryptocurrencies such as Bitcoin, Litecoin, Ethereum are always elusive. Hence, it would be a great value addition to investors if a model is able to predict what would be the nature of the crypto market for the next day. Through this paper, a time-series model using Long Short-Term Memory Networks is built to determine the value of cryptocurrency in the future. As a study, three cryptocurrencies – Bitcoin, Litecoin and Ethereum has been taken into consideration. A comparison of the results by using opinion mining to interpret the mood of the market on the current day for different currencies has been done. The sentiment scores got from natural language processing of textual data are used as features to the model used for predictions. The time-series charts are plotted using Plotly – python library for graphing plots. The Mean Absolute Error calculated between the actual and predicted values is used as the uncertainty quantification method. These uncertainty quantification methods are compared to analyze the present-day scenario of the market using opinion mining.
Cryptocurrencies, like Bitcoin, have become increasingly popular over the last decade. The price of Bitcoin has gone through several cycles of highs and lows. As a result, it is a widely discussed topic, especially on platforms like Twitter. Sentiment analysis is a research area of Natural Language Processing. It is used to determine whether the text is positive, negative, or neutral. Twitter tweets are more challenging to analyze when compared to other forms of text, due to the presence of irregular grammar, emoticons, and sarcasm. This study intends to analyze the effect of tweets on the stock price of Bitcoin. In order to study the effect, the sentiment associated with each tweet is calculated using VADER, and also the profession and follower countassociated with verified users who tweet about bitcoin is found. Following this, a model is trained and tested using a combined dataset of tweet related data and historical bitcoin price data. It was found that the sentiment of tweets does correlate with the shift in the price of bitcoin.
Analysis of any digital currency is performed for identifying and quantifying uncertainties, estimating their impact on results with real-time market value. Cryptocurrency, as an encrypted form of currencies which is used for shopping, investment, money transfer and in other purpose now days. Most popular use of Bitcoins are investment because its price was unexpectedly high in past few years (data shown in content of paper). In this paper prediction of Bitcoin close price by using the ARIMA model has been performed. The ARIMA model is found suitable for the prediction of bitcoin prices because this model is used for prediction of time series data. The forecast of future values is provided based on seasonality and trend present in the price data. In terms of visualizations, results are manifest by using R programming language. The obtained results are then compared with actual prices and percent mean error is calculated. The present mean error is found here less than 6% for most of the values.
Over the past few years, Bitcoin has been a topic of interest of many, from academic researchers to trade investors. Bitcoin is the first as well as the most popular cryptocurrency till date. Since its launch in 2009, it has become widely popular amongst various kinds of people for its trading system without the need of a third party and also due to high volatility of Bitcoin price. In this paper, we propose a suitable model that can predict the market price of Bitcoin best by applying a few statistical analysis. Our work is done on four year's bitcoin data from 2013 to 2017 based on time series approaches especially autoregressive integrated moving average (ARIMA) model and the work finally could acquire an accuracy of 90% for deciding volatility in weighted costs of bitcoin in the short run.
There has been much debate about whether returns on financial assets, such as stock returns or commodity returns, are predictable; however, few studies have investigated cryptocurrency return predictability. In this article we examine whether bitcoin returns are predictable by a large set of bitcoin price-based technical indicators. Specifically, we construct a classification tree-based model for return prediction using 124 technical indicators. We provide evidence that the proposed model has strong out-of-sample predictive power for narrow ranges of daily returns on bitcoin. This finding indicates that using big data and technical analysis can help predict bitcoin returns that are hardly driven by fundamentals.
Chih‐Hung Wu, Chih-Chiang Lu, Yu-Feng Ma, Ruei-Shan Lu
Long short-term memory (LSTM) networks are a state-of-the-art sequence learning in deep learning for time series forecasting. However, less study applied to financial time series forecasting especially in cryptocurrency prediction. Therefore, we propose a new forecasting framework with LSTM model to forecasting bitcoin daily price with two various LSTM models (conventional LSTM model and LSTM with AR(2) model). The performance of the proposed models are evaluated using daily bitcoin price data during 2018/1/1 to 2018/7/28 in total 208 records. The results confirmed the excellent forecasting accuracy of the proposed model with AR(2). The test mean squared error (MSE), root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE) for bitcoin price prediction, respectively. The our proposed LSTM with AR(2) model outperformed than conventional LSTM model. The contribution of this study is providing a new forecasting framework for bitcoin price prediction can overcome and improve the problem of input variables selection in LSTM without strict assumptions of data assumption. The results revealed its possible applicability in various cryptocurrencies prediction, industry instances such as medical data or financial time-series data.
Bitcoin is an established cryptographic digital currency whose value lays in the computational complexity rather than a physical commodity. Bitcoin is an open source software program with three aspects. (i) Peer-to-Peer networklow barrier entry; (ii) Mininginevitable concentration of power; (iii) Software upgrades. The nodes on the network follow a decentralized consensus for establishing the value of ledger and updating the blockchain which serves as a single source of truth for all transactions. As cryptocurrencies are developing more compelling utilities, creating ever faster and safer payment systems they are shifting the "money paradigm". Bitcoins are an evolution in money and provide a unique opportunity to forecast their price unlike the existing fiat currencies. The goal of this paper is to implement, train and evaluate several machine learning models in order to predict the price of the most popular cryptocurrency -Bitcoins. The various machine learning algorithms employed are -Linear Regression, K-Nearest Neighbors, Ridge Regression, Lasso Regression,
Blockchain technology shows significant results and huge potential for serving as an interweaving fabric that goes through every industry and market, allowing decentralized and secure value exchange, thus connecting our civilization like never before. The standard approach for asset value predictions is based on market analysis with an LSTM neural network. Blockchain technologies, however, give us access to vast amounts of public data, such as the executed transactions and the account balance distribution. We explore whether analyzing this data with modern Deep Leaning techniques results in higher accuracies than the standard approach. During a series of experiments on the Ethereum blockchain, we achieved $4$ times error reduction with blockchain data than an LSTM approach with trade volume data. By utilizing blockchain account distribution histograms, spatial dataset modeling, and a Convolutional architecture, the error was reduced further by 26\%. The proposed methodologies are implemented in an open source cryptocurrency prediction framework, allowing them to be used in other analysis contexts.
Blockchain is disrupting the banking industry and contributing to the increased big data in banking. However, there exists a gap in research and development into blockchain-ed big data in banking from an academic perspective, and this gap is expected to have a significant negative impact on the adoption and development of blockchain technology for banking. In hope of motivating more active engagement by academics, researchers and bankers alike, we present the most comprehensive review of the impact of blockchain in banking to date by summarizing the opportunities and challenges from a bankers perspective. In addition, we also discuss the impact that big data from blockchain will have on banking data analytics in future and show the increasing importance of filtering and signal extraction for the banking industry. Whilst there is evidence of selected banks adopting blockchain technology in isolation or small groups, we find the need for extensive research and development into several aspects of banking with blockchain to overcome the challenges which are currently hindering its adoption in banking across the globe.
A cryptocurrency is a digital asset designed to work as a medium of exchange that uses cryptography to secure its transactions, to control the creation of additional units, and to verify the transfer of assets.Cryptocurrencies are a type of digital currencies, alternative currencies and virtual currencies. Cryptocurrencies use decentralized control as opposed to centralized electronic money and central banking systems. The decentralized control of each cryptocurrency works through a blockchain, which is a public transaction database, functioning as a distributed ledger.Neural Networks field has many techniques to perform predictions. They are widely used to predict the future values of stock exchange indicators variables. In this paper we will try to use Artificial Neural Network to predict cryptocurrencies close prices, and we'll study the difference in price change with the normal stock exchanges.
The sentiment in Twitter about Bitcoin have direct or indirect influence on overall market value of the Bitcoin. This research is concerned with predicting the volatile price of Bitcoin by analyzing the sentiment in Twitter and to find the relation between them. The tweets of Bitcoin collected from different news account sources are classified to positive or negative sentiments. The obtained percentage of positive and negative tweets are feed to RNN model along with historical price to predict the new price for next time frame. The accuracy for sentiment classification of tweets in two class positive and negative is found to be 81.39 % and the overall price prediction accuracy using RNN is found to be 77.62%.
Main objective of this study is to develop investment portfolio with diversifications using two different assets. Modern portfolio theory develop investment portfolio to maximize expected return based on a given level of market risk. This study selected cryptocurreny (Bitcoin) and stock price (Petronas Gas Berhad) as the combination in developing investment portfolio. In this analysis, mean return for Bitcoin is 9.890 %. Meanwhile, the mean return for stock price of Petronas Gas Berhad is -0.496 %.The value of correlation is between two assets is -0.372. Result shows the portfolio risk can be reduced with the diversification approach for different assets. Therefore, findings of this study are important for assisting investors to maximize their return for given level of investment risk.