Determines the most relevant keywords from an article headline




A simple NPM package that determines the most relevant keywords in a headline by considering article context.


$ npm install headline-parser

Use the --save flag to include the module in your package.json


Require the parser

var headline_parser = require("headline-parser");

// Declare variables for your headline and article summary. These have been edited to provide a good example.

var headline = 'China successfully develops drone defense system';

var body = 'china has tested a self-developed laser defense system against small-scale low-altitude drones, state media said on Sunday. Reportedly, the drone defense is designed to destroy small-scale drones flying within an altitude of 500 meters and at speeds below 50 meters per second. In addition to the drone network, china has developed stealth jets and has built one aircraft carrier.';

// Find the most relevant keywords in the headline, sorted by number of appearances in the body text
var important_keywords = headline_parser.findKeywords(headline, body, 3);

// => Returns the top three occuring words [ 'china', 'drone', 'defense' ], with 'defense' appearing most often.


findKeywords() accepts four arguments, of which the last two are optional.

.findKeywords(headline, body [, n][, args]);
Argument nameDescriptionPermitted values
headlineHeadline of articleString
bodyContext from the article. May be the entire article body, or just a few sample sentences. The more context, the greater the accuracy of the parser.String
(optional) nNumber of top keywords desired. If left out, the parser will return all keywords sorted by relevance.Integer
(optional) argsTakes an object containing parameters for the keyword-extractor module used to pull keywords from the headline. Default is {language:"english", return_changed_case:true}Object (see docs)

Running tests

Install the development dependencies by running the following command:

$ npm install

To run tests:

$ npm test

How does it work?

It's pretty simple. This parser uses the keyword-extractor module to obtain keywords from a headline (all non-stopwords), then sorts those words by how many times each word appears in the article body provided. For example, this is a great tool to use with the Twitter API if you plan to search or stream tweets that relate to a specific news article.

Some things to note: The module will not count partial appearances of keywords, or compounded keywords. For instance, if one of your headline keywords is ['china'], then neither "China", "china's" or "Indochina" will be counted as an appearance of that keyword. Additionally, unless the args object is supplied with a return_changed_case: false parameter, the module will count only the lowercase appearances of the word.