KRİPTO PARA FİYATLARININ KLASİK VE YAPAY SİNİR AĞI MODELLERİ İLE TAHMİNİ
FORECASTING THE CRYPTOCURRENCY PRICE USING THE CLASSICAL AND ARTIFICIAL NEURAL NETWORKS MODELS

Summary/Abstract: Cryptocurrencies are increasing in importance. While to start with they were used only in virtual reality platforms for games, nowadays they are being used by many institutions and organisations as payment instruments. Against security risks, they are produced by an algorithm called blockchain. Predicting cryptocurrencies as accurately as possible is of great importance for investors and decision makers. In the scope of this study, the prices of the most-used four cryptocurrencies (Bitcoin, Ethereum, Ripple, Litecoin) are forecast. The analysis conducted is expanded by determining the break points of each series under investigation via the multiple break point test. For the purpose of predicting the concerned series, both classical time series models and three different artificial neural networks models are employed. In addition, the combining methods are carried out to improve the results. Among these random walk-dominated series, the best results are obtained with Bitcoin, which is the most widely known and has the highest trading volume, compared with other cryptocurrencies in the study.

• Details
• Contents