Used Python and unsupervised machine learning to categorize and group cryptocurrencies that exhibit similar historical price change patterns.
- Prepared the data. Used StandardScalar for data normalization.
- Found the Best Value for k Using the Original Scaled DataFrame.
- Clustered Cryptocurrencies with K-means Using the Original Scaled Data.
- Optimized Clusters with Principal Component Analysis.
- Found the Best Value for k Using the PCA Data
- Cluster Cryptocurrencies with K-means Using the PCA Data
What is the best value for k?
In both data sets 4 is the best k value.
What is the impact of using fewer features to cluster the data using K-Means?
The following observations are made for the clusters using reduced features:
- All clusters are more defined.
- Clusters 0 and 2 are packed together more tightly.
- Clusters 1 and 3 are a further distance away from Clusters 0,2 compared with the original data that had more features.