An introduction to ROC analysis
Pattern Recognition Letters - Special issue: ROC analysis in pattern recognition
A framework for unsupervised spam detection in social networking sites
ECIR'12 Proceedings of the 34th European conference on Advances in Information Retrieval
Community-based features for identifying spammers in online social networks
Proceedings of the 2013 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining
| Hi-index | 0.00 |
Content filtering is a popular approach to spam detection. It focuses on analysis of the message content to identify spam. In this paper, we evaluate the use of social network analysis measures to improve the performance of a content filtering model. By measuring the degree centrality of message transfer agents, we observed performance improvements for spam detection in repeated experiments; e.g. a 70% increase in the proportion of spam detected with a false positive rate of 0.1%. We were also able to use anomaly detection to identify mislabeled messages in a publicly available spam data set. Messages claiming unusually long paths between the sender's message transfer agent and the recipient's message transfer agent turned out to be spam.