The problem: Tracker identification is based on whitelists and lists of known trackers (see attached paper Introduction to being a Privacy Detective - Investigating and Comparing Potential Privacy Violations in Mobile Apps Using Forensic Methods)
The aim: Identify trackers from Mobile Application Network traffic w/o resorting to whitelists of known trackers (Finding features of different trackers from network flow) and create a model to identify trackers from network flow using Machine Learning Algorithms.
1. Do trackers lead to characteristic patterns of network traffic? Note: There could be more than one cluster of characteristic patterns (e.g. different categories of trackers).
2. Can these characteristic patterns be identified using pattern recognition methods?
3. Does this work even when the network traffic is encrypted?