Identifying Discriminating Age Groups for Online Predator Detection

Children are a major user group of today’s internet, mostly focusing on its social applications like social networking sites. Their online activities are not entirely without risk, as is demonstrated by the presence of online predators, who misuse its potential to search for and contact them with purposes in the area of sexual abuse. In this chapter we investigate the presence of these predators on Social Networking Sites and present a method of automatically detecting them. In this process we adopt a genetic algorithm approach that discovers a relation between the amount of friends on an online profile and the age of its owner, searching for statistically significant differences between known offenders and a control group of standard users. After the creation of this model, thresholds are calculated that classify users into danger categories. Experiments on actual police data sources show some promising results in countering this stealthy threat.