Using community question-and-answer corpora for question answering.

Abstract

Community Question Answering (CQA) services are defined as dedicated platforms for

users to respond to other users’ questions, resulting in the building of a community where

users share and interactively give ratings to questions and answers (Liu et al., 2008). CQA

services are emerging as a valuable information resource that is rich not only in the

expertise of the user community but also their interactions and insights. However, current

research efforts in CQA services failed to address how the heavy dependence on other

users’ participation can be alleviated. Moreover, scholarly inquiries have yet to dovetail

into a composite research stream where techniques gleaned from question answering (QA)

research could be exploited for automating CQA services by harnessing its information

richness.

The goal of this research was to propose an approach that adapts and applies

techniques from QA research and related fields to harness user-generated questions,

answers and related metadata from CQA corpora. The research question addressed to

achieve the research goal was: Given a newly posed question not found in a CQA corpus,

how can this corpus be harnessed to return high-quality answers? The two research

objectives addressed in this research were:

1. To match a newly posed question to those of a similar nature found in the corpus.

2. To select high-quality answers to the newly posed question from all candidate

answers for similar questions.