Recently we introduced Magnet.AI, Magnet Forensics’ machine learning algorithm that helps triage devices to identify whether there is content or data on the devices that could show intent to lure, or groom, children for illicit sexual activity.
There’s more to Magnet.AI than just a set of keywords, time and date ranges, or locations. Those elements are important, but they don’t tell the whole story. A new white paper, available later this month, tells what needs to go into a contextual chat analysis or classification solution like Magnet.AI—and how it can potentially save considerable time in an investigation.
Our latest white paper starts with an in-depth description of what child luring is (and isn’t), and why it’s difficult for other types of analytic solutions, such as named entity recognition and sentiment analysis, to pinpoint the nuances of a luring conversation.
We’ll also describe technology-facilitated child luring, and offer more specifics about where the evidence might reside. In addition, we’ll detail how machine learning, such as the model introduced with Magnet.AI, uses data science to overcome investigative challenges. We talk about why “accuracy” isn’t enough of a metric, and how to balance it with a second metric, “precision.” Finally, we show how to make use of contextual content analysis in a typical case workflow.