Human or automatic answers? A user's based study

Cristina González-Caro, Mari Carmen Marcos, Ricardo Baeza-Yates, Liliana Calderón-Benavides

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Understanding the different ways of the informationseeking behavior of users, and their implications is an important challenge of Web Information Services. In this paper we examine the suitability of Web Search Engines (SE) and Community Question-Answering (community QA) to satisfy the different expectations of users: queries expecting knowledge-based answers, and queries expecting subjective opinion-based answers. From this categorization, we conducted a user study of the results obtained from three Web SE, and three Communities QA. In this way, we found that SE result in a better alternative when the query is designed to get a uni-objective answer, while communities QA provide better results if the question gets multi-objective answers. On the other hand, when the query expects a subjective answer, none of these systems provides a very effective answer that can fulfill the expectations from the user. Knowing more about the type of queries that require human support would be useful to improve the user search experience in SE.

Original languageEnglish
Title of host publicationProceedings of the Latin American Web Conference, LA-WEB 2008
Pages101-107
Number of pages7
DOIs
StatePublished - 2008
Externally publishedYes
EventLatin American Web Conference, LA-WEB 2008 - Vila Velha, Espirito Santo, Brazil
Duration: 28 Oct 200830 Oct 2008

Publication series

NameProceedings of the Latin American Web Conference, LA-WEB 2008

Conference

ConferenceLatin American Web Conference, LA-WEB 2008
Country/TerritoryBrazil
CityVila Velha, Espirito Santo
Period28/10/0830/10/08

Fingerprint

Dive into the research topics of 'Human or automatic answers? A user's based study'. Together they form a unique fingerprint.

Cite this