Principal Component Analysis

The goal of supervised learning methods (regression and classification) is to predict outcome/response variable Y using a set of p features (X1, X2... Xp) measured on n observations. We train the machine on 'labeled' data to predict outcomes for unforeseen data.

Unsupervised learning is a set of tools (principle component analysis and clustering) intended to explore only a set of features (X1, X2... Xp) and to discover interesting things about these features. This is often performed as part of an exploratory data analysis.

The challenge of unsupervised learning is that is is more subjective than supervised learning, as there is no simple goal for the analysis such as prediction of a response.

Principle Component Analysis (PCA)

Visualize n observations with measurements on a set of p features as part of an exploratory data analysis. Do this by examining 2-dimensional scatterplots. PCA produces a low-dimensional representation of a dataset that contains as much variation as possible.