[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119231-en":3,"doc-seo-119231-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},119231,687197207639,"Asher","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Forecasting Digital Asset Return - An Application of Machine Learning Model - Research Article","Study identifies a machine learning model that addresses shortcomings of traditional statistical methods for forecasting Bitcoin prices, and specifies conditions for model suitability. Using a multivariate large dataset of Bitcoin prices and market microstructure variables, the research evaluates three approaches: double deep Q-learning, XGBoost, and ARFIMA-GARCH. Results indicate double deep Q-learning delivers superior returns and Sortino ratio, plus one-step-ahead sign forecasting even on synthetic data. Insights support practitioners and regulators in selecting an economically viable cryptocurrency return-forecasting model.","International Journal of Finance & Economics  \nRESEARCH ARTICLE  OPEN ACCESS   \nForecasting Digital Asset Return: An Application of Machine Learning Model  \nVito Ciciretti1 | Alberto Pallotta2 | Suman Lodh3  | P. K. Senyo4  | Monomita Nandy5   \n1Independent Researcher, Berlin, Germany | 2Middlesex University Business School, Middlesex University, London, UK | 3Kingston Business School, Kingston Hill Campus, Kingston Upon Thames, Kingston University London, Surrey, UK | 4Department of Decision and Risk, University of Southampton  \nBusiness School, University of Southampton, Southampton, UK | 5Brunel Business School, Brunel University of London, Uxbridge, UK Correspondence: Monomita Nandy ([monomita.nandy@brunel.ac.uk](monomita.nandy@brunel.ac.uk))  \nReceived: 21 March 2023 | Revised: 21 September 2024 | Accepted: 12 October 2024  \nKeywords: bitcoin | digital asset | double deep Q-learning | forecasting price | machine learning | reinforcement learning | time-series  \nABSTRACT  \nIn this study, we aim to identify the machine learning model that can overcome the limitations of traditional statistical modelling techniques in forecasting Bitcoin prices. Also, we outline the necessary conditions that make the model suitable. We draw on a multivariate large data set of Bitcoin prices and its market microstructure variables and apply three machine learning models, namely double deep Q-learning, XGBoost and ARFIMA-GARCH. The findings show that the double deep Q-learning model outperforms the others in terms of returns and Sortino ratio and is capable of one-step-ahead sign forecast of the returns even on synthetic data. These critical insights in forecasting literature will support practitioners and regulators to identify an economically viable cryptocurrency forecasting return model.  \n1 | Introduction  \nIn recent years, there has been growing interest in Bitcoin investment as the cryptocurrency gains global popularity and acceptance in some countries (Xie, Chen, and Hu 2020; Rehman, Asghar, and Kang 2020) . There are more than 81 million crypto wallets user across the world as of November 2022 (Statista 2021). The rapid evolution of Bitcoin trading over the past years has often raised concerns among investors in terms of overvaluation, overreaction, and irrational behaviour of the cryptocurrency prices (Amini et al. 2013; Borgards and Czudaj 2020; Corbet and Katsiampa 2020; Mattke et al. 2021) . Investors, market practitioners, and regulators have shown vigorous interest in understanding and explaining the movements of cryptocurrency prices in detail (Raimundo Junior et al. 2020; Signature Bank failure, March 2023) . Nevertheless, understanding the drivers of changes in cryptocurrency prices remains an open question as the application of econometric and statistical modelling has largely failed to adequately provide actionable  \ninsights in forecasting Bitcoin prices (Chen et al. 2021; Wang, Andreeva, and Martin-Barragan 2023) .  \nGiven that Bitcoin transactions generate large data sets that can provide critical insights, it is therefore important to explore if big data analytical tools such as machine learning could be useful solution to overcome the limitation in forecasting Bitcoin prices (Tofangchi et al. 2021) . In addition, cryptocurrencies like Bitcoin are less efficient when compared to the traditional financial assets (Al-Yahyaee, Mensi, and Yoon 2018), in the context of volatility. Even though, we observe a decrease in this volatility over the time, but the historical volatility of Bitcoin remains almost 10 times higher than gold and several conventional currencies (Bianchetti, Ricci, and Scaringi 2018) .  \nMoreover, Bitcoin, possess a combination of properties of other traditional financial and speculative asset and has a low correlation with other financial instruments traded in the financial market (Klein, Thu, and Walther 2018) . Thus, following  \nThis is an open access article under the terms of the Creative Commons Attr","cbCaioDKKdQg3MU7","https://ap.wps.com/l/cbCaioDKKdQg3MU7","pdf",10877467,1,18,"English","en",105,"# Introduction\n## Research question and motivation\n## Data and modelling approach","[{\"question\":\"Which machine learning models are compared for forecasting Bitcoin returns?\",\"answer\":\"The study compares double deep Q-learning, XGBoost, and ARFIMA-GARCH using multivariate Bitcoin price and market microstructure data.\"},{\"question\":\"What performance advantage does double deep Q-learning show?\",\"answer\":\"Double deep Q-learning outperforms the other models in terms of returns and Sortino ratio and can produce one-step-ahead sign forecasts even on synthetic data.\"},{\"question\":\"Why do the authors argue traditional statistical modelling is insufficient?\",\"answer\":\"They highlight limitations of econometric and statistical techniques in providing actionable insights for forecasting Bitcoin prices, motivating the use of big data and machine learning.\"}]","Forecasting Digital Asset Return - An Application of Machine Learning Model - Research Article | PDF",1785723205,45,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"forecasting-digital-asset-return-an-application-of-machine-learning-model-research-article","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/forecasting-digital-asset-return-an-application-of-machine-learning-model-research-article/119231/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Which machine learning models are compared for forecasting Bitcoin returns?","Question",{"text":75,"@type":76},"The study compares double deep Q-learning, XGBoost, and ARFIMA-GARCH using multivariate Bitcoin price and market microstructure data.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What performance advantage does double deep Q-learning show?",{"text":80,"@type":76},"Double deep Q-learning outperforms the other models in terms of returns and Sortino ratio and can produce one-step-ahead sign forecasts even on synthetic data.",{"name":82,"@type":73,"acceptedAnswer":83},"Why do the authors argue traditional statistical modelling is insufficient?",{"text":84,"@type":76},"They highlight limitations of econometric and statistical techniques in providing actionable insights for forecasting Bitcoin prices, motivating the use of big data and machine learning.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]