covectric

Covectric is a simple vector based search engine using cosine similarity and tf–idf methods for finding text similarity between documents.

Define "covectric"

Two documents are said to be covectric if they share at least one dimension (or term).

Install

1> npm install covectric

Usage

1//instantiate the model
2var covectric = require('covectric');
3var model = new covectric.Model();
4
5//populate the vector space and weight tokens based on term frequency
6var id = 1;
7['hello','hi','hi there','hey','hi to you'].forEach(function(t) {
8	model.upsertDocument(id++, t, t);
9});
10model.recomputeVectorBaseTokenWeights();
11
12//search the vector space
13
14var results = model.search("hi bob", 3);
15console.log("Search Results:", results);
16

Samples

See the samples directory in the repo for more examples of using Covectric

Reference

http://en.wikipedia.org/wiki/Vector_space_model

http://en.wikipedia.org/wiki/Cosine_similarity

http://en.wikipedia.org/wiki/Tf-idf

License

MIT Licensed