Useful concepts

PCA Principal Component Analyisis: Statistical procedure that uses an orthogonal transformation to convert a set of observations of possibly correlated variables into a set of values of linearly uncorrelated variables

Feature: Making a reference to a certain characteristic of a certain data point. It is some metadata that we will use for our analysis.

Mixed membership model: A grouping method where a certain data point can belong to several groups.

Latent Dirichlet Allocation: Is a generative statistical model that allows sets of observations to be explained by unobserved groups that explain why some parts of the data are similar

Similarity: A similarity measure or similarity function is a real-valued function that quantifies the similarity between two objects