A simple clustering algorithm using securities data. All the data have been generated randomly following a normal distribution.
The clusters n=3 where selected using dendrogram where the clusters' data inside each one is as similar as possible. For the K-means, they were selected using the "Elbow method".
In this example there are 3 main features of the data: volume, No. of trades and trading frequency. The features are reduced to two dimensions, each one important to determine de liquidity of a certain security.
The dimensionality reduction follows a simple PCA.