Classification of high dimensional data

1 min read

high dimensionality makes it harder to visualize the data and to build a model

Luckily, sometimes a few dimensions explain most of the variability

Here 96 dimensional feature vectors generated by Gaussian distributions are reduced to 3 dimensions using Principal Componenet Analysis

They are visualized and classified

This is also a comparison of full scikit-learn classification suit

let’s take a look at the colab notebook

September 24, 2018