Question processing for comparative and evaluative questions for business intelligence

Date of Publication

2010

Document Type

Bachelor's Thesis

Degree Name

Bachelor of Science in Computer Science

College

College of Computer Studies

Department/Unit

Computer Science

Thesis Adviser

Nathalie Rose Lim-Cheng

Defense Panel Member

Charibeth K. Cheng
Rachel Edita O. Roxas

Abstract/Summary

Business analysis assists companies in making decisions but business data are too massive, and computations using database queries are complicated. There is a need for an intuitive questions answering question (QA) system for business intelligence which answers questions in natural language. The research focuses on interpreting and answering comparative and evaluative questions, which involves semantic analysis, mapping of comparative and evaluative predicates with their corresponding quantifiable set of criteria for evaluation, processing the question representation, and generating the answer. Company transactions from business news are manually populated to form the business data, and predicates, templates, and other domain information are defined. The system supports different forms of questions like conjunctions, disjunctions, superlatives, and negations, represents at least 15 predicates including computation of trends, and provides question construction assistance, answer elaborations, and customizable domain information. The research resulted to a multi-domain QA framework with a QA system for the domain of pharmacy customized and tested. Invited experts saw the benefits and potential of the research and the system in many areas like decision making in internal affairs. The system is able to answer the questions with accuracy, although some ambiguous question constructs and support for discourse still have to be worked on. Integration with an information extraction component that will populate the domain data will make the system more useful.

Abstract Format

html

Language

English

Format

Print

Accession Number

TU18516

Shelf Location

Archives, The Learning Commons, 12F, Henry Sy Sr. Hall

Physical Description

1 v., various foliations ; 28 cm.

This document is currently not available here.

Share

COinS