Hiding inside the mutually-beneficial model of online crowdsourcing are malicious campaigns, which target manipulating search results or leaving fake reviews on the web. Crowdsourced manipulation reduces the quality and trustworthiness of online social media, threatening the security of cyberspace as a whole. To mitigate this problem, we developed a classification model which filters out malicious campaigns from nearly 450,000 campaigns on popular crowdsourcing platforms. We then presented this blacklist on a website, where parties adversely affected by malicious campaigns, such as targeted websites owners, legitimate workers, owners of the crowdsourcing platforms, can use this website as a tool to identify and moderate potential malicious campaigns from the web.
Worcester Polytechnic Institute
Major Qualifying Project
All authors have granted to WPI a nonexclusive royalty-free license to distribute copies of the work, subject to other agreements. Copyright is held by the author or authors, with all rights reserved, unless otherwise noted.