Evaluation of cryptocurrency investment with fuzzy logic in on-chain analysis

Recently, cryptocurrencies, a technology that does not need any authority or center, have been increasing the scrutiny of all dimensions with unprecedented interest. Especially with the Covid-19 pandemic, it has been observed that the active trading volume of cryptocurrencies has increased, and qualitative and quantitative applications of the studies have been determined in the literature in this context. For this reason, this study deals with the interpretation of mathematical transactions between two blocks of 4 crypto-assets (Bitcoin, Ethereum, Chainlink, Maker) that are actively traded in the market, in other words, on-chain analysis as a decision problem that has not yet been studied as far as is known from the literature. The proposed methodology includes the combination of the Random Forest Regression algorithm, which performs exceptionally in the prediction error of the data set with a series of comparisons of cryptocurrencies, and the Pythagorean fuzzy sets, which offer strong and large scale in the multi-criteria decision-making branch of decision science in fuzzy logic. In this conceptual framework, the fuzzy logic method is used while prioritizing the metrics that will affect the investment decision. In addition, the strong ranking of these metrics is achieved with the TOPSIS algorithm. In addition, sensitivity analyses to obtain stable rankings will provide an important projection to the literature by rationalizing the high uncertainty on the investors' side of crypto-assets traded in the markets. The results show that the Exchange Reserve metric of the crypto asset is the most important criterion among the stock market flow data, even when an investment preference is required for a crypto asset that is relatively new or whose historical data is not available for analysis.