Forecasting of cryptocurrency prices using artificial intelligence techniques

Bitcoin is the most popular cryptocurrency unit built on the block chain technology infrastructure. Under current circumstances, the future value of Bitcoin, which has the largest volume of the cryptocurrency market, is highly curious. In this study, the value of Bitcoin / TL price is predicted by using the dollar / TL rate and date attributes. The aim of the study is to provide a short and medium term options for both the users and the investors, by helping to make an accurate and timely investment in the price forecasting of Bitcoin, which is closely related to the financial sector and commercial transactions. Artificial intelligence methods such as artificial neural network (ANN), adaptive neuro fuzzy inference system (ANFIS), curve fitting and long short-term memory (LSTM) are used in the study. The reason of using multiple artificial intelligence and deep learning method in the study is to analyze the success rate of analysis of financial data of deep learning which has very high performance values and success rates especially in image processing and to compare the performance of the methods used. This study uses the data drawn from the web address of Koineks companies for getting BTC price and USD/TRY exchange rate are provided by the daily opening price value of the Central Bank of the Republic of Turkey. The date feature is based on the start and end dates as consecutive numerical numbers and covers the period between 08.12.2016 and 08.12.2018. As a result of the analyzes, the Bitcoin/TL price forecasting is evaluated as significant and successful by using the curve fitting and ANFIS methods. As a result of experimental studies, it is able to forecast the rate of other cryptocurrencies in real time on the web with the algorithms determined within the best parameters and time intervals. Keywords: Cryptocurrency, Bitcoin, Artificial Intelligent, Deep Learning, BTC Price Estimation, LSTM, ANFIS, Curve Fitting, ANN