PROJECT TITLE :

Demographic Information Inference through Meta-Data Analysis of Wi-Fi Traffic - 2018

ABSTRACT:

Privacy inference through meta-data (e.g., IP, Host) analysis of Wi-Fi traffic poses a probably more serious threat to user privacy. 1st, it provides a a lot of economical and scalable approach to infer users' sensitive data without checking the content of Wi-Fi traffic. Second, meta-information based demographics inference will work on both unencrypted and encrypted traffic (e.g., HTTPS traffic). In this study, we have a tendency to present a unique approach to infer user demographic data by exploiting the meta-information of Wi-Fi traffic. We tend to develop an inference framework based on machine learning and evaluate its performance on a true-world dataset, which includes the Wi-Fi access of twenty eight,158 users in 5 months. The framework extracts four types of options from real-world Wi-Fi traffic and applies a novel machine learning technique (XGBoost) to predict user demographics. Our analytical results show that, the overall accuracy of inferring gender and education level of users will be eighty two and 78 %, respectively. It is shocking to point out that, even for HTTPS traffic, user demographics will still be predicted at accuracy of 69 and seventy six p.c, respectively, that well demonstrates the practicality of the proposed privacy inference theme. Finally, we discuss and evaluate potential mitigation strategies for such inference attacks.