Using quantitative methods in cryptocurrency market daily value estimation

The aim of the study is to research which quantitative method gives more successful results in estimating the daily value of cryptocurrencies. In this context, 5-year daily data of Bitcoin, Ethereum, Binance Coin and Monero cryptocurrencies were analyzed with artificial neural networks, moving average, exponential smoothing, LSTM and RNN methods. The results were measured with MAPE, MSE and MAE error metrics. In addition, cointegration and causality tests were performed between cryptocurrencies. When the study results are evaluated according to the MAPE error metric, it is seen that artificial neural networks make the most successful predictions in Bitcoin prediction, exponential smoothing in Ethereum and Binance Coin prediction and the moving average method in Monero prediction. In addition, the daily values of the cryptocurrencies subject to the thesis study in 2024 were estimated with the artificial neural networks method and the relevant findings were presented in the appendix of the study. Keywords: Cryptocurrency Market, Artificial Neural Networks, Moving Averages Method, Exponential Smoothing Method, Deep Learning Methods.