|
Title:
|
CREDIBILITY MEASURE IN TWEET RETRIEVAL BASED
ON TEXTUAL CONSISTENCY |
|
Author(s):
|
Ibtissem Mejbri and Lobna Hlaoua |
|
ISBN:
|
978-989-8704-19-1 |
|
Editors:
|
Piet Kommers and Guo Chao Peng |
|
Year:
|
2020 |
|
Edition:
|
Single |
|
Keywords:
|
Credibility, Tweet Retrieval, Comments, Textual Consistency |
|
Type:
|
Full |
|
First Page:
|
189 |
|
Last Page:
|
196 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Nowadays, social media is a space where information can be shared concerning any type of event. These networks have
become a mass of social information that is accumulated day by day. This accumulation can even be doubled in a time of
sudden crisis or even within the spread of a pandemic as the example of the famous COVID-19 virus. This social
information can, without a doubt, impact, not only the internet surfer, but also the others who have no relationship with
the social networks. The major problem here is based on the credibility of information, since the users have the ability of
sharing any kind of information. This, in its way, explains why the credibility of social information presents an important
challenge. Consequently, in this paper, we are particularly interested in the social media Twitter, where our primary goal
is to evaluate the credibility score of a tweet. More precisely, we have presented a hybrid system that takes into
consideration, on one hand, the content and the author of the tweet, and on the other hand, includes the notion of textual
coherence between the tweet and its comments. In order to validate our proposed approach, we performed our work on a
database based on the tweets concerning the COVID-19. Experimental study shows the efficiency of our system, since it's
able to reach the same level of judgment as a human. The advantage of our approach based on adapting to changes, where
each comment added can also influence credibility, unlike most of the work which is based on machine learning. |
|
|
|
|
|
|
