A sample project to test different techniques to calculate the similarity between two sets.
Based on “Mining of Massive Datasets” by A.Rajaraman, J. Leskovec, J.D. Ullman, see http://www.mmds.org/
Utility that calculates Jaccard Index/Similarity between two collections.
The Jaccard similarity of sets S and T is |S ∩ T |/|S ∪ T |, that is, the ratio
of the size of the intersection of S and T to the size of their union.
See usage below, or check tests.
var left = new[] { 0,1,2,5,6 };
var right = new[] { 0,2,3,4,5,7,9 };
var index = JaccardIndexCalculator.GetIndex(left, right); // 0.33A helper class that implements MinHash algorithm for "documents", result can be used to calculate similarity.
Simple helper that extracts and return an array of strings that represents the source document as shingle-set.