Home // SOTICS 2019, The Ninth International Conference on Social Media Technologies, Communication, and Informatics // View article

Predicting Opinions Across Multiple Issues in Large Scale Cyber Argumentation Using Collaborative Filtering and Viewpoint Correlation

Authors:
Md Mahfuzer Rahman
Joseph Sirrianni
Xiaoqing (Frank) Liu
Douglas Adams

Keywords: opinion prediction; incomplete ongoing discussion; collaborative filtering; cyber argumentation; collective intelligence.

Abstract:
One challenging problem in large-scale cyber argumentation is that discussions are often incomplete as some ideas only get addressed by a fraction of the users. Typically, users engage only with some ideas but not all of them, making it difficult to assess collective intelligence. To resolve this problem, we developed an innovative method of predicting a user’s opinion on ideas that they have not discussed using the opinions from related ideas with intelligent argumentation and collaborative filtering. Our method considers the similarity of users and the correlation of different ideas across issues to make predictions. Compared to other existing opinion prediction methods, experimental results on an empirical dataset show that our method is 21.7% more accurate. Two major innovative contributions are made in this research: 1) We developed a novel approach to predict a participant’s opinion on a non-participated idea using similar users’ opinions from related ideas with an excellent accuracy in cyber argumentation; 2) This is the first research to enable multi issue opinion prediction with partial agreement on an idea. This is encouraging from several perspectives. This prediction model will help to assess collective intelligence from cyber argumentation more accurately by providing additional data both in individual and collective level. In addition, it may speed up a cyber argumentation analysis process by reducing the amount of participation required.

Pages: 45 to 51

Copyright: Copyright (c) IARIA, 2019

Publication date: November 24, 2019

Published in: conference

ISSN: 2326-9294

ISBN: 978-1-61208-757-3

Location: Valencia, Spain

Dates: from November 24, 2019 to November 28, 2019