Herding behavior in exploring the predictability of price clustering in cryptocurrency market

Abstract

We propose the K-means approach and the Hidden Markov Model (HMM), to predict the phenomenon of price clustering of the cryptocurrency market. This approach aims to understand the relationship of price clustering with herding behavior, volatility, price, and economic policy uncertainty (EPU). Our results indicate that the (HMM) with four states has the best one-step-ahead forecasting performance. The results gave new insights into the financial analysis of cryptocurrency market about the dynamic relationship between price clustering regimes and different states of the explanatory variable. Our finding proves the efficiency of (HMM) for our sample and provides a good predictability.