Robust outliers detection method for ethereum exchange rate: a statistical approach using high frequency data

Ethereum is one of the cryptocurrency that attracts attention from investors in year of 2017. Therefore, the objective of this study is to evaluate the data distribution of Ethereum exchange rate to validate the dynamic behavior of price movement. The finding of this study will help investors to understand the volatility of Ethereum exchange rate data. This study implemented Shapiro-Wilk normality test including graphical test to detect the outliers in the exchange rate data. The p-value of Shapiro-Wilk test is 0.0000. This value indicates the distribution of first difference for Ethereum exchange rate is not a normal distribution data. Then, this finding is validated using histogram and normal percentiles plot. Both of this plots indicates non-normal distribution because the data distribution does not follow normal distribution reference line. Finally, Box-Whisker plot is performed to detect the existence of outliers in the data. Result indicates there are suspected outliers and outliers in the Ethereum exchange rate data. This concluded that first difference of Ethereum exchange rate data is highly volatile. The important finding from this study is the dynamic behavior of Ethereum exchange rate is highly volatile and high risk. Therefore, any investors that interested with Ethereum cryptocurrency need to monitor closely the price to prevent high loss of their investmen