國立高雄大學圖資館 |

Language: English

Back

Network Analysis and Modeling for Ne...

State University of New York at Stony Brook.

Network Analysis and Modeling for News and Social Media.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Network Analysis and Modeling for News and Social Media.
Author:	Ye, Junting.
Published:	Ann Arbor : ProQuest Dissertations & Theses, 2019
Description:	124 p.
Notes:	Source: Dissertations Abstracts International, Volume: 81-04, Section: B.
Notes:	Advisor: Skiena, Steven.
Contained By:	Dissertations Abstracts International81-04B.
Subject:	Computer science.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13879624
ISBN:	9781687999221

Network Analysis and Modeling for News and Social Media.
Ye, Junting.

Network Analysis and Modeling for News and Social Media. - Ann Arbor : ProQuest Dissertations & Theses, 2019 - 124 p.

Source: Dissertations Abstracts International, Volume: 81-04, Section: B.

Thesis (Ph.D.)--State University of New York at Stony Brook, 2019.

This item must not be sold to any third party vendors.

Network modeling plays important roles in solving real-world problems, e.g. webpage ranking on hyperlink graph, information diffusion on social network, fraud transaction detection on consumer-merchant network, etc.This thesis focuses on designing network analysis and modeling methods, and their applications in online news and social media problems. More specifically, (i) I formulate the problem of robust semi-supervised learning on multiple graphs and propose a novel algorithm to learn graph weights. (ii) Fake reviews are prevalent nowadays. Observing distinct network localities between legitimate users and opinion spammers, a new method is designed to detect spammers in review networks from e-commerce platforms like Amazon and Itunes. (iii) In recent years, fake news and growing distrust in traditional news media challenges the foundation of democracy. I conduct the first in-depth study on large-scale news citation graph and find it an important signal to characterize peer reputation in journalism. I also design and implement a news ranking system, MediaRank, which tracks over 50,000 online news sources around the world (www.media-rank.com). (iv) People tend to communicate with others of similar background, a.k.a. communication homophily. Based on this observation, I propose a method to learn distributed representation (namely name embeddings) for first and last names. These embeddings are learned from real-world communication networks, including email (non-public data) and social media (public data). Extensive experiments prove that name embeddings achieves superior performances against traditional name substring in encoding gender, ethnic and nationality signals. NamePrism, an online API system for ethnicity and nationality classification, has been supporting over 150 research groups from social science, political science, etc. (www.name-prism.com).

ISBN: 9781687999221Subjects--Topical Terms:

199325
Computer science.

Network Analysis and Modeling for News and Social Media.
LDR:02841nmm a2200289 4500 001 570764
005 20200514111954.5
008 200901s2019 ||||||||||||||||| ||eng d
020 $a 9781687999221
035 $a (MiAaPQ)AAI13879624
035 $a AAI13879624
040 $a MiAaPQ $c MiAaPQ
100 1 $a Ye, Junting. $3 857454
245 1 0 $a Network Analysis and Modeling for News and Social Media.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2019
300 $a 124 p.
500 $a Source: Dissertations Abstracts International, Volume: 81-04, Section: B.
500 $a Advisor: Skiena, Steven.
502 $a Thesis (Ph.D.)--State University of New York at Stony Brook, 2019.
506 $a This item must not be sold to any third party vendors.
520 $a Network modeling plays important roles in solving real-world problems, e.g. webpage ranking on hyperlink graph, information diffusion on social network, fraud transaction detection on consumer-merchant network, etc.This thesis focuses on designing network analysis and modeling methods, and their applications in online news and social media problems. More specifically, (i) I formulate the problem of robust semi-supervised learning on multiple graphs and propose a novel algorithm to learn graph weights. (ii) Fake reviews are prevalent nowadays. Observing distinct network localities between legitimate users and opinion spammers, a new method is designed to detect spammers in review networks from e-commerce platforms like Amazon and Itunes. (iii) In recent years, fake news and growing distrust in traditional news media challenges the foundation of democracy. I conduct the first in-depth study on large-scale news citation graph and find it an important signal to characterize peer reputation in journalism. I also design and implement a news ranking system, MediaRank, which tracks over 50,000 online news sources around the world (www.media-rank.com). (iv) People tend to communicate with others of similar background, a.k.a. communication homophily. Based on this observation, I propose a method to learn distributed representation (namely name embeddings) for first and last names. These embeddings are learned from real-world communication networks, including email (non-public data) and social media (public data). Extensive experiments prove that name embeddings achieves superior performances against traditional name substring in encoding gender, ethnic and nationality signals. NamePrism, an online API system for ethnicity and nationality classification, has been supporting over 150 research groups from social science, political science, etc. (www.name-prism.com).
590 $a School code: 0771.
650 4 $a Computer science. $3 199325
690 $a 0984
710 2 $a State University of New York at Stony Brook. $b Computer Science. $3 708806
773 0 $t Dissertations Abstracts International $g 81-04B.
790 $a 0771
791 $a Ph.D.
792 $a 2019
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13879624