Published Paper Details:

Cryptocurrency is becoming more and more important in changing the financial system as a result of its rising adoption by merchants and popularity. Despite the fact that many people are investing in cryptocurrencies, it is still difficult to forecast their dynamical characteristics, unpredictability, and predictability, which raises the risk of the investment. Understanding the factors that affect the creation of value is essential. Using cutting-edge artificial intelligence frameworks like as fully connected Artificial Neural Networks (ANN) and Long Short-Term Memory (LSTM) Recurrent Neural Networks, we assess the price dynamics of Bitcoin, Ethereum, and Ripple in this study. We find that LSTM is more effective than ANN at leveraging significant information concealed in historical memory long-term history because it tends to depend more on short-term dynamics. However, given sufficient historical data, ANN can attain comparable precision as LSTM. This study gives a novel illustration of the predictability of the Cryptocurrency market price. However, the explanation for predictability may differ depending on the type of the machine-learning model employed.