Feature Extraction
A feature is synonymous to input a variable or attribute. Feature extraction is very different from Feature selection: the former consists in transforming arbitrary data, such as text or images, into numerical features usable for machine learning. Feature extraction involves reducing the amount of resources required to describe a large set of data, and is very important in with respect to Big Data. When we perform analysis of large, complex data sets, we are dealing with thousands of variables, which may seemingly have no rhyme nor rhythm. I heard somewhere, though I cannot remember where or from whom, that typically only 40 features are considered when making predictive models. The data I work with has 50 times that number, and feature extraction can be time consuming and tedious using traditional means.