Developing a Machine Learning Model to Catch Sexual Predators
- 0 Collaborators
The FBI estimates that more than 750,000 sexual predators are online at any given moment. Sextortion may be perpetrated by people already known to the victim, and may involve interactions by mobile phone, text, or email. We've used existing pre-classified datasets of phishing emails and aggressive texts from online support forums to develop sample machine learning models to detect potential phishing attempts or aggressive language, and these have been implemented as live running demos on our website using Python and OpenWhisk. This project intends to build on this work by using a larger dataset obtained from scraping text conversations from the web in cases that led to convictions relating to sexual discussions with persons they believed to be underage. This project supports our entry into the IBM Watson AI XPRIZE competition. ...learn more
Overview / Usage
The FBI estimates that more than 750,000 sexual predators are online at any given moment. Sextortion may be perpetrated by people already known to the victim, and may involve interactions by mobile phone, text, or email.
We've used existing pre-classified datasets of phishing emails and aggressive texts from online support forums to develop sample machine learning models to detect potential phishing attempts or aggressive language, and these have been implemented as live running demos on our website using Python and OpenWhisk.
This project intends to build on this work by using a larger dataset obtained from scraping text conversations from the web in cases that led to convictions relating to sexual discussions with persons they believed to be underage.
This project supports our entry into the IBM Watson AI XPRIZE competition.