Clustering and Portfolio Managing with a Mathematical Model for Cryptocurrencies

The objectives of this research study are (1) clustering cryptocurrencies (2) constructing an investment portfolio according to Markowitz Model via RStudio and Excel Solver. The data are collected from the CCI30 index from June 1, 2020, to October 1, 2021, in which the cryptocurrency market is continuously growth. The daily data used to analyze in this research are open, low, high and closed prices. From this study, it is found that cryptocurrencies can be categorized into 4 main groups which are (1) high volatility and high return (2) low volatility but high return (3) high volatility but low return (4) high speculation. These results of the cryptocurrency grouping are used to construct the portfolio according to Markowitz's portfolio theory. Three types of portfolios are considered: (1) Minimum-Variance Portfolio (2) Maximum-Sharpe Ratio (3) MaximumExpected-Return. By considering three groups of cryptocurrencies for constructing portfolio: (1) low risk group, (2) high return per risk group (3) high return group, the results reveal that, for low risk group and high return per risk group, the constructed portfolios are able to generate more return than that of the market while the high return group of cryptocurrencies, only the Minimum-Variance Portfolio can generate the better return