Survey on modern predictive models for cryptocurrency price prediction

Cryptocurrency markets have become a volatile and dynamically changing field that has attracted researchers and investors’ attention over the last few years. Huge fluctuations in cryptocurrency prices show an urgent need for accurate forecasting models. The successful price prediction in this field is very complex as it features many elements such as market sentiment, macro trends. In this paper a thorough review is given for the methodologies used in Cryptocurrency which includes approaches such as Machine Learning (ML), Deep Learning (DL), Hybrid models. Through this paper, we have shed light on Long-Short-Term Memory (LSTM), Hybrid models, and Sentiment analysis. This paper combines relevant literature for highlighting important insights and future research direction which can lead to more accurate models for price prediction, which will smoothen the part of investment decisions in volatile market.