Method of Fraudster Fingerprint Formation During Mobile Application Installations

Abstract

This study aimed to develop a method of fraudster fingerprint formation during mobile application installations, based on a fuzzy model for fraudster fingerprint formation and algorithms of its use for fraudster fingerprint formation. This method allows determining the reason of labeling user by a particular class during fraud detection. The use of the developed method in fraud detection tasks makes it possible to correctly identify 99.56% of users in general and 80.43% of correctly determined fraudsters in particular and speed up the fraud detection process.