Closed MontyCoder0701 closed 1 year ago
The initial clustering looks pretty bad. We can try to improve the data preprocessing, change the hyperparameters, or try another clustering algorithm. Also there are not enough datasets to begin with.
Changed the data preprocessing to standard scaler, which did improve the cluster.
https://medium.com/swlh/k-means-clustering-on-high-dimensional-data-d2151e1a4240 Apply PCA to reduce to two dimensions
https://www.pythonforfinance.net/2018/02/08/stock-clusters-using-k-means-algorithm-in-python/ https://coinmarketcap.com/tokens/views/all/
Cluster the tokens and see what clusters show up.