Question

Which technique implicitly defines the class of possible patterns by introducing a notion of similarity between data?

Answer 1

The answer is class of kernel methods.

The class of kernel methods implicitly defines the class of possible patterns by introducing a notion of similarity between data.

The similarity between documents could be based on any feature such as length, language or topic to mention just a few. The choice of similarity depends on the choice of the feature chosen.

Answer 2

Kernel methods are quite useful in pattern analysis. This software especially for machine learning is used to find and study different types of relations between clusters, correlation, classifications and all different types of patterns available in the data sets. 

These kernel methods are widely used for sequencing data, for graphs, text, images and even vectors. Kernel methods thus allow a close demonstration of linear relations in high dimensional spaces at a very cost.