vector space model
Table of Contents
In the context of Representation Learning,Text Representation.., feature engineering involves representing data numerically for Machine Learning algorithms to be able to process the said data.
1. Properties
Some useful features of vector spaces that can employed in devising algorithms (non-exhaustive):
- The notion of distance between two vectors
- The notion of similarity between two vectors (cosine, etc)
- Orthogonality to denote independence
- projections onto hyperplanes
- maps to higher dimensional spaces
2. Vector Visualization
Visualizing vectors w.r.t each other in a set might provide important insights regarding the dataset (could be images, text, speech, etc) that they represent.