國立虎尾科技大學 |

Analysis of User Generated Content in Digital Platforms.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Analysis of User Generated Content in Digital Platforms./
Author:	Tripathi, Sambit.
Description:	1 online resource (126 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-03, Section: A.
Contained By:	Dissertations Abstracts International85-03A.
Subject:	Web studies. -
Online resource:	click for full text (PQDT)
ISBN:	9798380322195

Analysis of User Generated Content in Digital Platforms.
Tripathi, Sambit.

Analysis of User Generated Content in Digital Platforms. - 1 online resource (126 pages)

Source: Dissertations Abstracts International, Volume: 85-03, Section: A.

Thesis (Ph.D.)--University of Massachusetts Lowell, 2023.

Includes bibliographical references

This dissertation focuses on applying data analysis methods to user-generated content (UGC) produced on digital platforms. User-generated content (UGC) is a unique attribute of the Internet that influences individuals' or organizations' behavior on digital platforms. Some examples of UGC are retail transactions, social media posts or activities, online reviews, job posts, and video content. Social media, e-commerce, online labor markets, and other digital platforms rely on both the generation and consumption of UGC to survive in the industry.The first essay (Chapter 1) focuses on the evolution of text attributes associated with online reviews on e-commerce platforms. Using a large dataset, we extract sentiments, along with novel attributes - product usage contexts and product features - present in each online review and analyze their pattern over the order of the review. Results indicate that the amount of sentiment intensity, product features, and usage contexts diminish concerning the increase in review order, suggesting that the earlier part of the reviews tends to have higher product information for prospective customers. The second essay (Chapter 2) focuses on peer endorsements among workers in online labor markets. Online labor markets connect buyers with gig workers across several task categories. A buyer evaluates workers' quality based on their past performance encapsulated in ratings and reviews. However, these ratings can be inflated and arguably fail to assess workers' true quality. In this study, we explore how gig workers' personality traits in terms of Social Value Orientation (SVO) affect their performance in an online labor market. We measure SVO from peer endorsements among workers in an online labor market. Our results show that a cooperative SVO, where gig workers endorse each other, is more beneficial to the stakeholders of online labor markets than a competitive and individualistic SVO. We also explore how such cooperative behavior evolves by leveraging social network analysis methods to examine the endorsements generated among gig workers. We observe that reciprocity, homophily, and worker popularity induce such cooperative behavior. In the third essay (Chapter 3), we develop an explainable retail item recommendation approach using topic modeling and word embeddings on historical retail transactions of an online retail platform. Online retailing platforms are currently using neural network-based recommender systems to recommend relevant products to their users based on the user's purchase history. Although these systems show performance improvement over traditional recommender systems, it is often difficult to explain why and how the recommended items are provided to specific users. In this study, we develop a novel method that exploits the latent theme and intent behind users' transaction histories to explain the recommended items generated by a neural embedding model. Our proposed method explains the recommended items based on the purchased items that share the same theme as the recommended items. We have compared our method against state-of-the-art baseline methods and found that our method achieves improved explanations over the baseline methods.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798380322195Subjects--Topical Terms:

1148502
Web studies.
Subjects--Index Terms:

Digital platformsIndex Terms--Genre/Form:

554714
Electronic books.

Analysis of User Generated Content in Digital Platforms.
LDR:04649ntm a22004097 4500 001 1151345
005 20241104085611.5
006 m o d
007 cr bn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798380322195
035 $a (MiAaPQ)AAI30569881
035 $a AAI30569881
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Tripathi, Sambit. $3 1478093
245 1 0 $a Analysis of User Generated Content in Digital Platforms.
264 0 $c 2023
300 $a 1 online resource (126 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-03, Section: A.
500 $a Advisor: Deokar, Amit;Li, Xiaobai.
502 $a Thesis (Ph.D.)--University of Massachusetts Lowell, 2023.
504 $a Includes bibliographical references
520 $a This dissertation focuses on applying data analysis methods to user-generated content (UGC) produced on digital platforms. User-generated content (UGC) is a unique attribute of the Internet that influences individuals' or organizations' behavior on digital platforms. Some examples of UGC are retail transactions, social media posts or activities, online reviews, job posts, and video content. Social media, e-commerce, online labor markets, and other digital platforms rely on both the generation and consumption of UGC to survive in the industry.The first essay (Chapter 1) focuses on the evolution of text attributes associated with online reviews on e-commerce platforms. Using a large dataset, we extract sentiments, along with novel attributes - product usage contexts and product features - present in each online review and analyze their pattern over the order of the review. Results indicate that the amount of sentiment intensity, product features, and usage contexts diminish concerning the increase in review order, suggesting that the earlier part of the reviews tends to have higher product information for prospective customers. The second essay (Chapter 2) focuses on peer endorsements among workers in online labor markets. Online labor markets connect buyers with gig workers across several task categories. A buyer evaluates workers' quality based on their past performance encapsulated in ratings and reviews. However, these ratings can be inflated and arguably fail to assess workers' true quality. In this study, we explore how gig workers' personality traits in terms of Social Value Orientation (SVO) affect their performance in an online labor market. We measure SVO from peer endorsements among workers in an online labor market. Our results show that a cooperative SVO, where gig workers endorse each other, is more beneficial to the stakeholders of online labor markets than a competitive and individualistic SVO. We also explore how such cooperative behavior evolves by leveraging social network analysis methods to examine the endorsements generated among gig workers. We observe that reciprocity, homophily, and worker popularity induce such cooperative behavior. In the third essay (Chapter 3), we develop an explainable retail item recommendation approach using topic modeling and word embeddings on historical retail transactions of an online retail platform. Online retailing platforms are currently using neural network-based recommender systems to recommend relevant products to their users based on the user's purchase history. Although these systems show performance improvement over traditional recommender systems, it is often difficult to explain why and how the recommended items are provided to specific users. In this study, we develop a novel method that exploits the latent theme and intent behind users' transaction histories to explain the recommended items generated by a neural embedding model. Our proposed method explains the recommended items based on the purchased items that share the same theme as the recommended items. We have compared our method against state-of-the-art baseline methods and found that our method achieves improved explanations over the baseline methods.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Web studies. $3 1148502
653 $a Digital platforms
653 $a Online reviews
653 $a Peer endorsements
653 $a Social Value Orientation
653 $a Text mining
653 $a User generated content
655 7 $a Electronic books. $2 local $3 554714
690 $a 0310
690 $a 0454
690 $a 0646
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a University of Massachusetts Lowell. $b Information Systems. $3 1478094
773 0 $t Dissertations Abstracts International $g 85-03A.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30569881 $z click for full text (PQDT)