Building a sentiment corpus using a gamified framework

Date of Publication


Document Type

Master's Thesis

Degree Name

Master of Science in Computer Science


College of Computer Studies


Computer Science

Thesis Adviser

Solomon See

Defense Panel Chair

Charibeth Cheng

Defense Panel Member

Ralph Vincent Regalado


Gamification, the use of game mechanics and game elements in non-game contexts, is an emerging approach for crowdsourcing the collection of data. This study uses a gamified application in the form of an online debate game as a cost-efficient way to build an agreement-objection corpus from which a sentiment corpus can potentially be derived. This approach has advantages over traditional ways which are difficult, time-consuming, and expensive, since there is currently no automated way to determine sentiment polarity. It allows for a time-efficient and cost-efficient way of building a sentiment corpus that can be applied to several natural language processing tasks and research.

Polarity, the gamified application used for this study, was able to collect 626 statements after being deployed for 43 days. The cleaning process was able to filter out 72.88% of the noise data. After the cleaning process, a total of 596 statements remained in the agreement-objection corpus. Assuming that this corpus can be mapped into a sentiment corpus by assuming all agree statements are positive and everything else are negative, it results into a sentiment corpus of 82.86% accuracy.

Abstract Format






Accession Number


Shelf Location

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

Physical Description

leaves ; 4 3/4 in.

This document is currently not available here.