An Ensemble Learning Framework for Online Web Spam Detection

• Published in: 2013 12th International Conference on Machine Learning and Applications Vol. 1; pp. 40 - 45
• Main Authors: Cailing Dong, Bin Zhou
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2013
• Subjects: Browsers; Detectors; ensemble learning; Labeling; online web spam detection; personalization; Search engines; Servers; Unsolicited electronic mail; Web pages
• DOI: 10.1109/ICMLA.2013.15

Most of the existing studies about web spam detection explicitly or implicitly assume that the detection process is performed offline on the search engine side. However, we argue that online web spam detection is even useful in some specific scenarios. We propose to implement a web browser plug-in to support online web spam detection. Three different sets of spam labeling data are collected and adopted for learning a reliable web spam classifier. An empirical study is conducted on the benchmark web spam data collection. The statistical analysis of the data set verifies the necessity of online web spam detection. The performance of the proposed ensemble learning framework for online web spam detection is also examined and it meets the requirement of online webs Pam detection.