Authors:
Costin Chiru
;
Traian Rebedea
and
Adriana Erbaru
Affiliation:
University Politehnica of Bucharest, Romania
Keyword(s):
CSCL, Natural Language Processing, Participant Assessment, PageRank, Online Conversations.
Related
Ontology
Subjects/Areas/Topics:
Communities of Practice
;
Computer-Supported Education
;
Enterprise Information Systems
;
Learning/Teaching Methodologies and Assessment
;
Social Media Analytics
;
Society, e-Business and e-Government
;
Software Agents and Internet Computing
;
Web 2.0 and Social Networking Controls
;
Web Information Systems and Technologies
Abstract:
In this paper we present a novel approach that uses the well-known PageRank algorithm for assessing multi-threaded chat conversations. As online conversations can be modelled as directed graphs, we have investigated a method for allowing a real-time analysis of the conversation using PageRank by computing the ranks of the utterances based on the explicit and implicit links available in the discussion. This model has been also extended to offer a method for computing connections between the debated topics and the chat participants and between each of the debated topics in the conversation, called the participant-topic and the topic-topic attraction. The results presented in this paper are promising, but also reflect several important differences between the existent offline analysis tools for chats and the PageRank method.