Ranking of tweets based on credibility factors / Pradeesh

Twitter is extensively being used to share news, links, images and even have conversations. In Malaysia alone, there are 3.5 million twitter users. As the volume of tweets and users who are increasingly accessing tweets as source of information, they have less information to judge if a tweet is cred...

詳細記述

書誌詳細
第一著者: Pradeesh, -
フォーマット: 学位論文
出版事項: 2017
主題:
_version_ 1849735856843128832
author Pradeesh, -
author_facet Pradeesh, -
author_sort Pradeesh, -
description Twitter is extensively being used to share news, links, images and even have conversations. In Malaysia alone, there are 3.5 million twitter users. As the volume of tweets and users who are increasingly accessing tweets as source of information, they have less information to judge if a tweet is credible or not. The consequences of spreading non-credible tweets can be harmful to the society, nation and to the entire world. To respond to this issue, this research considered ranking tweets by various qualities of a tweet, such as popularity, reliability, timeliness, trustworthiness of web pages and tweets link to provide a more credible Twitter users search results than the current Twitter search which only looks at relevance without looking at the credibility of the tweet. An evaluation of the method on 144,972 tweets from GST which is consists of Malay and English tweets shows that the proposed scoring technique pTRank scores much more better compared to TwitterRank in various ranking evaluations such as in Normalized Discounted Cumulative Gain (nDCG), the system scored a score of 0.393, as opposed to TwitterRank which is at 0.121. The same trend is also noticed with both GST tweets in both the languages and as well as only on English.
format Thesis
id oai:studentsrepo.um.edu.my:14261
institution Universiti Malaya
publishDate 2017
record_format eprints
spelling oai:studentsrepo.um.edu.my:142612023-04-10T22:40:58Z Ranking of tweets based on credibility factors / Pradeesh Pradeesh, - QA75 Electronic computers. Computer science ZA Information resources Twitter is extensively being used to share news, links, images and even have conversations. In Malaysia alone, there are 3.5 million twitter users. As the volume of tweets and users who are increasingly accessing tweets as source of information, they have less information to judge if a tweet is credible or not. The consequences of spreading non-credible tweets can be harmful to the society, nation and to the entire world. To respond to this issue, this research considered ranking tweets by various qualities of a tweet, such as popularity, reliability, timeliness, trustworthiness of web pages and tweets link to provide a more credible Twitter users search results than the current Twitter search which only looks at relevance without looking at the credibility of the tweet. An evaluation of the method on 144,972 tweets from GST which is consists of Malay and English tweets shows that the proposed scoring technique pTRank scores much more better compared to TwitterRank in various ranking evaluations such as in Normalized Discounted Cumulative Gain (nDCG), the system scored a score of 0.393, as opposed to TwitterRank which is at 0.121. The same trend is also noticed with both GST tweets in both the languages and as well as only on English. 2017-05 Thesis NonPeerReviewed application/pdf http://studentsrepo.um.edu.my/14261/2/Pradeesh.pdf application/pdf http://studentsrepo.um.edu.my/14261/1/Pradeesh.pdf Pradeesh, - (2017) Ranking of tweets based on credibility factors / Pradeesh. Masters thesis, Universiti Malaya. http://studentsrepo.um.edu.my/14261/
spellingShingle QA75 Electronic computers. Computer science
ZA Information resources
Pradeesh, -
Ranking of tweets based on credibility factors / Pradeesh
title Ranking of tweets based on credibility factors / Pradeesh
title_full Ranking of tweets based on credibility factors / Pradeesh
title_fullStr Ranking of tweets based on credibility factors / Pradeesh
title_full_unstemmed Ranking of tweets based on credibility factors / Pradeesh
title_short Ranking of tweets based on credibility factors / Pradeesh
title_sort ranking of tweets based on credibility factors pradeesh
topic QA75 Electronic computers. Computer science
ZA Information resources
url-record http://studentsrepo.um.edu.my/14261/
work_keys_str_mv AT pradeesh rankingoftweetsbasedoncredibilityfactorspradeesh