Browse Prior Art Database

Method and System for Automatically Refining a Set of Search Results using Natural Language Processing

IP.com Disclosure Number: IPCOM000239536D
Publication Date: 2014-Nov-13
Document File: 3 page(s) / 31K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method and system is disclosed for automatically refining a set of search results using Natural Language Processing (NLP). The method and system reads a first set of results obtained from a first set of search terms to automatically build a second set of search terms for retrieving a refined set of search results.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 43% of the total text.

Page 01 of 3

Method and System for Automatically Refining a Set of Search Results using Natural Language Processing

Disclosed is a method and system for automatically refining a set of search results using Natural Language Processing (NLP).

The method and system reads a first set of search results returned from a search using a first set of search terms. The method and system, then, automatically builds a second set of search terms based on a machine reading of the first set of search results to provide a second set of search terms. Thereafter, the method and system sends out a query to a search engine based on the second set of search terms to retrieve the refined set of search results. The method and system, then, presents both the first set of search terms and results and the second set of search terms and results according to one or more user preferences.

In an embodiment, the method and system provides a search selection capability to a user via a user interface for presenting search results. The search selection capability can either show only the refined set of search results or show the refined set of search results side by side the first set of search results.

In accordance with the embodiment, the method and system enables the user to select a number of articles that can be machine read. The method and system stores an initial set of keywords obtained from reading the articles in at least one of an array, and a hash table and structure capable of being sorted and otherwise manipulated through an Application Programming Interface (API). Once the user sets the search selection capability, the method and system submits the initial set of keywords to a search site. Thereafter, the method and system returns the first set of search results and presents them on the left side of a side by side viewing window.

The method and system, then, machine reads and interprets the content of the first set of search results. For example, if the method and system is configured by the user to read first ten articles of the first set of search results, the method and system reads only the ten articles and applies NLP to the ten articles for a deeper understanding of the content. NLP can include techniques such as, but not be limited to, one or more statistical inference algorithms and deep parsing in order to understand the root of an article message, a simple count of a number of times a certain word comes up and how many articles can be used to derive important keywords that can be used in a follow-on search. The one or more statistical interference algorithms can also take into account a possible relationship between original keywords provided and terms appearing in the articles.

The method and system statistically ranks a list of phrases from the first set of search results in accordance with a broader subject related to the first set of search terms to identify the second set of search terms. The statistical ranking can be based on, but not limited t...