Browse Prior Art Database

Question Formation Enhancement

IP.com Disclosure Number: IPCOM000223766D
Publication Date: 2012-Nov-28
Document File: 1 page(s) / 26K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to optimize question and answer systems so that a user receives suggestions for questions.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 56% of the total text.

Page 01 of 1

Question Formation Enhancement

Question and Answer (Q/A) systems work best when a well formed question is crafted and submitted to the system. The well-formed question leads to the highest quality results. However the act of building a well-formed question is an art, rather than a prescriptive science. If one asks a question, one gets back the answer as a reflection on the formed question, in example, a question crafted with only acronyms.

The disclosed method optimizes the crafting of a question through the process of retrieving the set of question-answer pairs and the related satisfaction with the previous use of the pair. The method then generates from the set a model by first, observing the grammar, taxonomy of the domain, language, natural language parsing of the pair. The method uses the model to predict the effectiveness of the question through the generated model and present the associated confidence level of the question. The model is then used to suggest, by the rules of the model, modifications to the question such that a higher confidence level is achieved.

The method may be iteratively applied to suggest improvements as the question is refined.

An exemplary series of steps to operate the system would be:
User logs into Q/A System.

User types question (eg "What is BOB?") System analyzes the question

The system extracts natural language parsing to analyze the grammar, taxonomy of the domain, user and language.

The question may be analyzed within a sco...