Browse Prior Art Database

Computational Method to Determine Best Matches from Search Results

IP.com Disclosure Number: IPCOM000244339D
Publication Date: 2015-Dec-03
Document File: 7 page(s) / 137K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to automatically compute and determine the best matches from referenced documents (e.g., values, data items (columns), and data set (table) in search results), and then use those in a formal query or join operation. This identifies the values that best match the intention of the query.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 01 of 7

Computational Method to Determine Best Matches from Search Results


In the Business Analytics industry, many products strive to provide features that allow users to enter a natural language question,

which the system then parses, analyzes, and automatically maps to a formal query, such as Structured Query Language (SQL), or uses to automatically build a star schema model. Users often ask questions related member values in datasets or tables, which are indexed by a search engine. The system either fully or partially matches the words in the question to member values of one or more columns. However, the user cannot easily identify which values best match the intention of the query.

Currently, the user determines the relevance cutoff in search results. Automatically selecting the best matches and using the results in a formal query or join operation, requires a computational approach to find the best matches.

The novel contribution is a method to automatically compute and determine the best matches from referenced documents (e.g., values, data items (columns), and data set (table) in search results), and then use those in a formal query or join operation.

The following example embodiment illustrates the steps for implementing the computational method to determine best matches from search results.

A movie producer(i.e. user) wants to analyze the box office revenues of some movies, using a dataset of Hollywood movies. The user enters Query 1: "box office of X Men *".

For X Men , the search engine returns 10 search results partially matched to X Men .

Figure 1: Search results partially matched to X Men

1


Page 02 of 7

The user can easily see only the first six results are relevant X-Men movies, but the search engine does not know without the user's input.

The novel computational method can determine that the first six results are the best matches , and then automatically add that text as the filter of a formal query.

The user enters Query 2: "box off...