Bitcoin is the world's first decentralized digital crypto currency which does not need an intermediary like a bank and is most secure because of block chain implementation. The price of a single bitcoin has been increasing drastically since 2010 as a form of digital gold. Thus, bitcoin is very volatile as its price changes every second which is a high risk for investors. The purpose of this paper is to analyse the machine learning algorithms which are of maximum efficiency in predicting the bitcoin price. I have explored many machine learning regression-based algorithms to build a prediction model for analysing future bitcoin prices. This paper is based on a deep learning-based artificial neural network model named GRU (Gated Recurrent Unit) to predict bitcoin future prices accurately based on past price information available. Root Mean Square Error and Mean Absolute Percent Error are the key performance indicators to measure forecast accuracy.
Rane Nikita, S J Subhashini, B S Yashwin, Vishal K Vavle · 5 authors
Cryptocurrency is a computerized method of money where transactions are done in the cyber space. This is a digital/soft currency, unlike currency notes they are not hard copies. Emphasizing the varieties of currencies that are decentralized and do not have any third party involvement, therefore ensuring that users can get all the services. Because of the high volatility the currency impacts the international trade and relations. Etherium, Ripple, Bitcoin, Litecoin some examples of popular currencies that exists. The study on popular cryptocurrencies Bitcoin, Ripple, Etherium are performed every year. An effective way to improve the method of predictions using deep learning models namely Grated Reccurrent Unit (GRU) and Long Short-Term Memory (LSTM). The research is devoted to the problems related to predicting crypto currency prices using machine learning and data science. The main algorithms used are: RNN and GRU. The data set will include the past price information from the popular crypto currencies for example: Bitcoin, Ethereum and Ripple. LSTM, RNN and GRU algorithms already exist to predict the future price action but the predictability rate is subpar. The main goal is to combine RNN and GRU Algorithms to form a hybrid and possibly increase the accuracy of the predictions.
With the advancement of blockchain technology and the development of digital economy, more investors are entering the cryptocurrency market, and the use of historical information as a means to evaluate and forecast future trends in the rapidly changing cryptocurrency market has become a major topic at the moment. Based on the SVR model, this paper proposes a cryptocurrency price expectation model based on long-term and short-term integrated learning and uses a large amount of historical cryptocurrency price data to analyze and verify the integrated learning model. Experimental results indicate that the accuracy of the SVR model for cryptocurrency price prediction can be effectively improved by the integrated learning model.
We start by presenting a short description of the concept of cryptocurrency and the history behind it. Recently-developed literature that attempt to predict volatilities of cryptocurrency valuations through creation of hybrid artificial neural network models are then discussed. For the major part of the paper, we delve into details of multiple hybrid artificial neural networks that were thoroughly implemented to predict cryptocurrency volatilities. Results are reported within the form of a survey. Finally, we compare different methods and discuss their results follow at the end.
Deep Reinforcement Learning solutions have been applied to different control problems with outperforming and promising results. In this research work we have applied Proximal Policy Optimization, Soft Actor-Critic and Generative Adversarial Imitation Learning to strategy design problem of three cryptocurrency markets. Our input data includes price data and technical indicators. We have implemented a Gym environment based on cryptocurrency markets to be used with the algorithms. Our test results on unseen data shows a great potential for this approach in helping investors with an expert system to exploit the market and gain profit. Our highest gain for an unseen 66 day span is 4850 US dollars per 10000 US dollars investment. We also discuss on how a specific hyperparameter in the environment design can be used to adjust risk in the generated strategies.
Abstract Since the emergence of Bitcoin, cryptocurrencies have grown significantly, not only in terms of capitalization but also in number. Consequently, the cryptocurrency market can be a conducive arena for investors, as it offers many opportunities. However, it is difficult to understand. This study aims to describe, summarize, and segment the main trends of the entire cryptocurrency market in 2018, using data analysis tools. Accordingly, we propose a new clustering-based methodology that provides complementary views of the financial behavior of cryptocurrencies, and one that looks for associations between the clustering results, and other factors that are not involved in clustering. Particularly, the methodology involves applying three different partitional clustering algorithms, where each of them use a different representation for cryptocurrencies, namely, yearly mean, and standard deviation of the returns, distribution of returns that have not been applied to financial markets previously, and the time series of returns. Because each representation provides a different outlook of the market, we also examine the integration of the three clustering results, to obtain a fine-grained analysis of the main trends of the market. In conclusion, we analyze the association of the clustering results with other descriptive features of cryptocurrencies, including the age, technological attributes, and financial ratios derived from them. This will help to enhance the profiling of the clusters with additional descriptive insights, and to find associations with other variables. Consequently, this study describes the whole market based on graphical information, and a scalable methodology that can be reproduced by investors who want to understand the main trends in the market quickly, and those that look for cryptocurrencies with different financial performance.In our analysis of the 2018 and 2019 for extended period, we found that the market can be typically segmented in few clusters (five or less), and even considering the intersections, the 6 more populations account for 75% of the market. Regarding the associations between the clusters and descriptive features, we find associations between some clusters with volume, market capitalization, and some financial ratios, which could be explored in future research.
Financial data are expensive and highly sensitive with limited access. We aim to generate abundant datasets given the original prices while preserving the original statistical features. We introduce the Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) into the field of the stock market, futures market and cryptocurrency market. We train our model on various datasets, including the Hong Kong stock market, Hang Seng Index Composite stocks, precious metal futures contracts listed on the Chicago Mercantile Exchange and Japan Exchange Group, and cryptocurrency spots and perpetual contracts on Binance at various minute-level intervals. We quantify the difference of generated results (836,280 data points) and original data by MAE, MSE, RMSE and K-S distances. Results show that WGAN-GP can simulate assets prices and show the potential of a market simulator for trading analysis. We might be the first to look into multi-asset classes in a systematic approach with minute intervals across stocks, futures and cryptocurrency markets. We also contribute to quantitative analysis methodology for generated and original price data quality.
Sezercan TANIŞMAN, Abdullah Ammar Karcıoğlu, Aybars Uğur, Hasan Bulut
Finansal varlıkların gelecekteki değerlerinin tahmini yatırımcılar için varlıklarını korumak adına önemlidir. 2008 yılında hayatımıza giren ve finansal varlıklar konusunda radikal bir değişiklik olan Bitcoin ise eski ve yeni yatırımcıların ilgisini çekmiş durumdadır. Ancak Bitcoin, doğası gereği diğer finansal varlıklara göre değerini belirleyen farklı parametreler içermektedir ve geleneksel tahmin yöntemleri Bitcoin gibi çok hareketli değerlere sahip finansal varlıkları tahmin etmekte güçlük çekmektedir. Bu çalışmada çok değişkenli LSTM sinir ağı ve klasik ARIMA zaman serisi modeli kullanılarak Bitcoin’in gelecek değerinin tahmini için modeller geliştirilmiştir. Uygulanan iki modelin tahmin doğruluğu performans değerlendirme metrikleri olan hata metrikleri kullanılarak karşılaştırılmıştır. Deneysel çalışmalar sonucu, LSTM sinir ağı modeli yakın ve uzak gelecek için düşük hata oranı ile tahmin performansı gerçekleştirirken ARIMA zaman serisi modeli yakın gelecek tahmini için düşük hata oranı ile tahmin performansı gerçekleştirmiştir.
In equity trading, internalization is the predominant execution method for uninformed order flow, allowing retail brokers to realize cost savings and thereby offer price improvements to customers. In cryptocurrency trading, there are doubts as to whether informed and uninformed traders can be distinguished in the same way, leading brokers to seek cost savings through internal order matching instead. Using the historical order flow of the German cryptocurrency broker BISON, we present a prediction-based approach to internal order matching: Upon receiving a customer order, our model forecasts whether future order flow will be sufficient to neutralize the order before the settlement date. With a prediction accuracy of 85%, it enables brokers to match three-quarters of order volume internally, which is three times as much as a traditional static approach, and realize meaningful cost savings, even after accounting for common minimum price improvements.
Bitcoin is a high-risk asset with a potentially high return. Predicting Bitcoin candlestick, i.e., open, high, low, and close (OHLC) prices, can help investors make trading decisions. The objective of this study is to develop a neural network model to predict the candlestick prices of Bitcoin for the next period. Additionally, this study investigates methods to enhance the model's forecasting performance by feature transformations, specifically data normalization. This study employs two neural network algorithms, Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), to forecast daily Bitcoin OHLC prices. To enhance the model's performance, we compare sliding window normalization with whole set normalization techniques. The normalization techniques investigated for both whole set and sliding window data include z-score normalization, min-max normalization, and relative change normalization. Furthermore, this study compares two candlestick prediction methods, namely using OHLC prices and using candle wick (CULR) to predict OHLC prices. The models use historical OHLC prices over several days to predict the next day's OHLC prices. The results indicate that the best-performing model is the OHLC method using GRU algorithm with sliding window z-score normalization, which achieves an MAPE of 1.95% and an RMSE of 767.71. Moreover, the sliding window normalization generally outperforms the whole set normalization for both LSTM and GRU models in terms of RMSE and MAPE. Regarding the candlestick prediction methods, there was no significant difference in their performance in terms of accuracy and forecasting error. However, our results suggest that the OHLC method performs slightly better than the CULR method.
Sowmya Dunnala, Anusha Bandla, Krishna Sai Anjana Sunkara, Ebenezer Jangam
Bitcoin is the most trending cryptocurrency which is used worldwide. Nowadays many general people or investors investing on bitcoin. But it becomes great challenge to analyze or predict the bitcoin price. Because of its fluctuations it is very hard to predict the price of the bitcoin. By this time machine learning came into picture with many models to analyze the behavior of bitcoin price by using time series data. These models will give better insights to the people who wants to invest on the bitcoin and they will able to understand about the volatility of bitcoin. We can use many machine learning models for prediction. But accuracy of the model is the deciding factor. We used ARIMA, LSTM and Facebook Prophet models and after the prediction is over, we have designed an ensemble model which merges the different models. And based upon the error rate we have decided the best model.
Nowadays, digital cryptocurrencies are the most popular asset, especially for international exchanges. Bitcoin is the earliest cryptocurrency that succeeded in being used in financial transactions. Bitcoin stores the transactions in Blockchain technology. Bitcoin price has been unstable during the time from 0.5$ to about 60,000$ since 2010. Many efforts exist to predict Bitcoin value or its fluctuations using machine learning techniques. The price prediction is usually more challenging than fluctuations prediction, and its performance metrics are improved. This study introduces a methodology to predict Bitcoin price in a dataset, including four intervals to evaluate the proposed method in different situations. The experimental results show that the generalized linear model and Long Short-Term Memory (LSTM) were the best machine learning techniques. The proposed model outperforms the deep learning baseline model with about 18% and 20% relative improvement in mean absolute error and means absolute percentage error, respectively. Deep learning approaches have achieved much better results than other approaches due to the automatic selection of features. Compared to the results reported in the literature, the 1D-CNN+IndRNN proposed approach has reached 81% accuracy, with an 18% improvement. In the proposed approach, 1D-CNN is responsible for feature extraction and IndRNN is responsible for learning features in the form of time series.