<![CDATA[This study aims to (1) classify cryptocurrencies based on the integration of market data and on-chain data to provide a more accurate mapping of digital asset characteristics. The method used is Multi-Channel Clustering, which enables the combination of multiple data views (multi-view) in the clustering process. The data includes market capitalization, trading volume, percentage gain, and volatility from the market data channel, as well as coin supply and active addresses from the on-chain channel. All data were normalized using the Min-Max Normalization method to ensure scale uniformity across variables. The clustering process was carried out using a K-Means algorithm adapted for the multi-channel context. The results of the study identified three main clusters: Cluster 1 contains coins with medium to low market and on-chain activity characteristics such as Tron and Cro; Cluster 2 includes coins with high volume and significant on-chain activity but not dominant, such as Ethereum and Solana; and Cluster 3 consists solely of Bitcoin, which has a unique profile in both channels. In conclusion, the Multi-Channel Clustering method proves effective in producing a more comprehensive classification of cryptocurrencies and can serve as a decision-support tool in highly volatile and complex market environments. Keywords: Cryptocurrancy, Classification, Multi-Channel Clustering.]]>
Copyrights © 2025
