{"id":{"repo_id":"ohiolink","oai_identifier":"oai:etd.ohiolink.edu:osu1365520334"},"canonical_url":"https://search.dev.ndltd.org/etd/ohiolink/oai:etd.ohiolink.edu:osu1365520334","repository":{"repo_id":"ohiolink","name":"OhioLINK","base_url":"https://etd.ohiolink.edu/acprod/odb_etd/ws/oai/oai"},"display":{"title":"Spam Analysis and Detection for User Generated Content in Online Social Networks","abstract":"Recent years have witnessed the success of a number of online social networks(OSNs) and explosive increasing of social media. These social networking andsocial media sites have attracted a significant number of participants thatcontribute various types of contents on the Internet, which are generallyreferred as user generated content (UGC). A well designed UGC network canutilize the wisdom of crowds to collect, organize, and vote user contributedcontent to generate high quality knowledge with a relatively low cost. However,the open environment of UGC system also makes it easy to be polluted andattacked by spammers and malicious users. How users participate in UGCnetworks, especially how users contribute content and share content with theirfriends and other users, is fundamental to spam detection and high qualityknowledge discovery. In this dissertation, we investigate two importantresearch issues: (1) discovering user content generation patterns in OSNs,focusing on publicly available content (knowledge sharing), and (2) detectingspam in user generated content based on our discovered patterns.With the access to three large OSN user activity logs, including Yahoo! Blogs,Yahoo! Answers, and Yahoo! Del.icio.us, for a duration of up to 4.5 years, weare able to well analyze the patterns of content generation patterns of socialnetwork users in detail. Our analysis consistently shows that users' postingbehavior in these networks exhibits strong daily and weekly patterns, but theuser active time in these OSNs does not follow commonly assumed exponentialdistributions. We also show that the user posting behavior in these OSNsfollows stretched exponential distributions instead of widely accepted powerlaw distributions. Our discovery lays a foundation for user behavior analysisin social networks, and serves as a ground truth for anomaly detection andanti-spam.Applying the user posting behavior distribution pattern, we further conducted acomprehensive analysis of spamming activities on a large commercial social blogUGC site in 325 days covering over 6 million posts and nearly 400 thousandusers. Observing power law distribution instead of our discovered stretchedexponential distribution on user contributions, we find it actually indicatesserious UGC spam attack activities. Our analysis shows that UGC spammersexhibit unique non-textual patterns, such as posting activities, advertisedspam link metrics, and spam hosting behaviors. Based on these non-textualfeatures, we show with commonly used classification methods that a highdetection rate could be achieved offline. These results further motivate us todevelop a runtime scheme, BARS, to detect spam posts based on these spammingpatterns. The experimental results demonstrate the effectiveness androbustness of BARS.To timely detect spam in large social network sites, it is desirable todiscover self-tuned, unsupervised schemes that can save the training cost ofsupervised classification schemes. Identifying the limitations of existingunsupervised detection schemes due to assumptions of spammer behaviors that nolonger hold, we design an unsupervised spam detection scheme, called UNIK.Instead of picking out spammers directly, UNIK leverages both theconnection-based social graph and the content-based user-link graph to removenon-spammers from the network first, and then clusters spammers with thelanding pages they are trying to advertise. Based on highly accurate detectionresults of UNIK, we further analyze a number of spam campaigns. The resultshows that different spammer clusters demonstrate distinct characteristics,implying the ability of UNIK to automatically extract spam signatures.","abstract_html":"Recent years have witnessed the success of a number of online social networks(OSNs) and explosive increasing of social media. These social networking andsocial media sites have attracted a significant number of participants thatcontribute various types of contents on the Internet, which are generallyreferred as user generated content (UGC). A well designed UGC network canutilize the wisdom of crowds to collect, organize, and vote user contributedcontent to generate high quality knowledge with a relatively low cost. However,the open environment of UGC system also makes it easy to be polluted andattacked by spammers and malicious users. How users participate in UGCnetworks, especially how users contribute content and share content with theirfriends and other users, is fundamental to spam detection and high qualityknowledge discovery. In this dissertation, we investigate two importantresearch issues: (1) discovering user content generation patterns in OSNs,focusing on publicly available content (knowledge sharing), and (2) detectingspam in user generated content based on our discovered patterns.With the access to three large OSN user activity logs, including Yahoo! Blogs,Yahoo! Answers, and Yahoo! Del.icio.us, for a duration of up to 4.5 years, weare able to well analyze the patterns of content generation patterns of socialnetwork users in detail. Our analysis consistently shows that users' postingbehavior in these networks exhibits strong daily and weekly patterns, but theuser active time in these OSNs does not follow commonly assumed exponentialdistributions. We also show that the user posting behavior in these OSNsfollows stretched exponential distributions instead of widely accepted powerlaw distributions. Our discovery lays a foundation for user behavior analysisin social networks, and serves as a ground truth for anomaly detection andanti-spam.Applying the user posting behavior distribution pattern, we further conducted acomprehensive analysis of spamming activities on a large commercial social blogUGC site in 325 days covering over 6 million posts and nearly 400 thousandusers. Observing power law distribution instead of our discovered stretchedexponential distribution on user contributions, we find it actually indicatesserious UGC spam attack activities. Our analysis shows that UGC spammersexhibit unique non-textual patterns, such as posting activities, advertisedspam link metrics, and spam hosting behaviors. Based on these non-textualfeatures, we show with commonly used classification methods that a highdetection rate could be achieved offline. These results further motivate us todevelop a runtime scheme, BARS, to detect spam posts based on these spammingpatterns. The experimental results demonstrate the effectiveness androbustness of BARS.To timely detect spam in large social network sites, it is desirable todiscover self-tuned, unsupervised schemes that can save the training cost ofsupervised classification schemes. Identifying the limitations of existingunsupervised detection schemes due to assumptions of spammer behaviors that nolonger hold, we design an unsupervised spam detection scheme, called UNIK.Instead of picking out spammers directly, UNIK leverages both theconnection-based social graph and the content-based user-link graph to removenon-spammers from the network first, and then clusters spammers with thelanding pages they are trying to advertise. Based on highly accurate detectionresults of UNIK, we further analyze a number of spam campaigns. The resultshows that different spammer clusters demonstrate distinct characteristics,implying the ability of UNIK to automatically extract spam signatures.","abstract_has_math":false,"creators":["Tan, Enhua"],"institution":"The Ohio State University","degree_name":"Doctor of Philosophy","degree_level":"doctoral","degree_discipline":"Computer Science and Engineering","degree_department":null,"school":null,"contributors":["Zhang, Xiaodong"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2013,"date_issued":"2013-07-23","date_published":"2013-07-23","updated_at":"2026-07-24T03:37:31Z","subjects":["Computer Engineering","Computer Science","user generated content","online social networks","user behavior","stretched exponential distribution","spam filtering","spam detection","spam classification","decision tree","social graph","user-link graph","Sybil attack","community detection","BARS","UNIK"],"languages":["English"],"rights":["unrestricted","This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws."],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://rave.ohiolink.edu/etdc/view?acc_num=osu1365520334","outbound_label":"Repository record","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Zhang, Xiaodong"]},{"key":"dc:creator","label":"Author","values":["Tan, Enhua"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2013-07-23"]},{"key":"dc:publisher","label":"Institution","values":["The Ohio State University / OhioLINK"]},{"key":"dc:type","label":"Dc Type","values":["Electronic Thesis or Dissertation"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science and Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["The Ohio State University"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Computer Engineering","Computer Science","user generated content","online social networks","user behavior","stretched exponential distribution","spam filtering","spam detection","spam classification","decision tree","social graph","user-link graph","Sybil attack","community detection","BARS","UNIK"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["English"]},{"key":"dc:rights","label":"Dc Rights","values":["unrestricted","This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://rave.ohiolink.edu/etdc/view?acc_num=osu1365520334"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Recent years have witnessed the success of a number of online social networks(OSNs) and explosive increasing of social media. These social networking andsocial media sites have attracted a significant number of participants thatcontribute various types of contents on the Internet, which are generallyreferred as user generated content (UGC). A well designed UGC network canutilize the wisdom of crowds to collect, organize, and vote user contributedcontent to generate high quality knowledge with a relatively low cost. However,the open environment of UGC system also makes it easy to be polluted andattacked by spammers and malicious users. How users participate in UGCnetworks, especially how users contribute content and share content with theirfriends and other users, is fundamental to spam detection and high qualityknowledge discovery. In this dissertation, we investigate two importantresearch issues: (1) discovering user content generation patterns in OSNs,focusing on publicly available content (knowledge sharing), and (2) detectingspam in user generated content based on our discovered patterns.With the access to three large OSN user activity logs, including Yahoo! Blogs,Yahoo! Answers, and Yahoo! Del.icio.us, for a duration of up to 4.5 years, weare able to well analyze the patterns of content generation patterns of socialnetwork users in detail. Our analysis consistently shows that users' postingbehavior in these networks exhibits strong daily and weekly patterns, but theuser active time in these OSNs does not follow commonly assumed exponentialdistributions. We also show that the user posting behavior in these OSNsfollows stretched exponential distributions instead of widely accepted powerlaw distributions. Our discovery lays a foundation for user behavior analysisin social networks, and serves as a ground truth for anomaly detection andanti-spam.Applying the user posting behavior distribution pattern, we further conducted acomprehensive analysis of spamming activities on a large commercial social blogUGC site in 325 days covering over 6 million posts and nearly 400 thousandusers. Observing power law distribution instead of our discovered stretchedexponential distribution on user contributions, we find it actually indicatesserious UGC spam attack activities. Our analysis shows that UGC spammersexhibit unique non-textual patterns, such as posting activities, advertisedspam link metrics, and spam hosting behaviors. Based on these non-textualfeatures, we show with commonly used classification methods that a highdetection rate could be achieved offline. These results further motivate us todevelop a runtime scheme, BARS, to detect spam posts based on these spammingpatterns. The experimental results demonstrate the effectiveness androbustness of BARS.To timely detect spam in large social network sites, it is desirable todiscover self-tuned, unsupervised schemes that can save the training cost ofsupervised classification schemes. Identifying the limitations of existingunsupervised detection schemes due to assumptions of spammer behaviors that nolonger hold, we design an unsupervised spam detection scheme, called UNIK.Instead of picking out spammers directly, UNIK leverages both theconnection-based social graph and the content-based user-link graph to removenon-spammers from the network first, and then clusters spammers with thelanding pages they are trying to advertise. Based on highly accurate detectionresults of UNIK, we further analyze a number of spam campaigns. The resultshows that different spammer clusters demonstrate distinct characteristics,implying the ability of UNIK to automatically extract spam signatures."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf","p.131","5.73 MB"]},{"key":"dc:title","label":"Title","values":["Spam Analysis and Detection for User Generated Content in Online Social Networks"]}]}],"canonical_facts":{"dc:contributor":["Zhang, Xiaodong"],"dc:creator":["Tan, Enhua"],"dc:date":["2013-07-23"],"dc:description":["Recent years have witnessed the success of a number of online social networks(OSNs) and explosive increasing of social media. These social networking andsocial media sites have attracted a significant number of participants thatcontribute various types of contents on the Internet, which are generallyreferred as user generated content (UGC). A well designed UGC network canutilize the wisdom of crowds to collect, organize, and vote user contributedcontent to generate high quality knowledge with a relatively low cost. However,the open environment of UGC system also makes it easy to be polluted andattacked by spammers and malicious users. How users participate in UGCnetworks, especially how users contribute content and share content with theirfriends and other users, is fundamental to spam detection and high qualityknowledge discovery. In this dissertation, we investigate two importantresearch issues: (1) discovering user content generation patterns in OSNs,focusing on publicly available content (knowledge sharing), and (2) detectingspam in user generated content based on our discovered patterns.With the access to three large OSN user activity logs, including Yahoo! Blogs,Yahoo! Answers, and Yahoo! Del.icio.us, for a duration of up to 4.5 years, weare able to well analyze the patterns of content generation patterns of socialnetwork users in detail. Our analysis consistently shows that users' postingbehavior in these networks exhibits strong daily and weekly patterns, but theuser active time in these OSNs does not follow commonly assumed exponentialdistributions. We also show that the user posting behavior in these OSNsfollows stretched exponential distributions instead of widely accepted powerlaw distributions. Our discovery lays a foundation for user behavior analysisin social networks, and serves as a ground truth for anomaly detection andanti-spam.Applying the user posting behavior distribution pattern, we further conducted acomprehensive analysis of spamming activities on a large commercial social blogUGC site in 325 days covering over 6 million posts and nearly 400 thousandusers. Observing power law distribution instead of our discovered stretchedexponential distribution on user contributions, we find it actually indicatesserious UGC spam attack activities. Our analysis shows that UGC spammersexhibit unique non-textual patterns, such as posting activities, advertisedspam link metrics, and spam hosting behaviors. Based on these non-textualfeatures, we show with commonly used classification methods that a highdetection rate could be achieved offline. These results further motivate us todevelop a runtime scheme, BARS, to detect spam posts based on these spammingpatterns. The experimental results demonstrate the effectiveness androbustness of BARS.To timely detect spam in large social network sites, it is desirable todiscover self-tuned, unsupervised schemes that can save the training cost ofsupervised classification schemes. Identifying the limitations of existingunsupervised detection schemes due to assumptions of spammer behaviors that nolonger hold, we design an unsupervised spam detection scheme, called UNIK.Instead of picking out spammers directly, UNIK leverages both theconnection-based social graph and the content-based user-link graph to removenon-spammers from the network first, and then clusters spammers with thelanding pages they are trying to advertise. Based on highly accurate detectionresults of UNIK, we further analyze a number of spam campaigns. The resultshows that different spammer clusters demonstrate distinct characteristics,implying the ability of UNIK to automatically extract spam signatures."],"dc:format":["application/pdf","p.131","5.73 MB"],"dc:identifier":["http://rave.ohiolink.edu/etdc/view?acc_num=osu1365520334"],"dc:language":["English"],"dc:publisher":["The Ohio State University / OhioLINK"],"dc:rights":["unrestricted","This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws."],"dc:subject":["Computer Engineering","Computer Science","user generated content","online social networks","user behavior","stretched exponential distribution","spam filtering","spam detection","spam classification","decision tree","social graph","user-link graph","Sybil attack","community detection","BARS","UNIK"],"dc:title":["Spam Analysis and Detection for User Generated Content in Online Social Networks"],"dc:type":["Electronic Thesis or Dissertation"],"thesis:degree_discipline":["Computer Science and Engineering"],"thesis:degree_level":["doctoral"],"thesis:degree_name":["Doctor of Philosophy"],"thesis:institution_name":["The Ohio State University"]},"updated_at":"2026-07-24T03:37:31Z"}