Evaluating Machine Learning Techniques for Smart Home Device Classification
Abstract
Smart devices in the internet of things have transformed the management of personal and industrial spaces. Recent research has shown that it is possible to model a subjects pattern-of-life through Bluetooth Low Energy (BLE) and Wi-Fi smart device data leakage. A key step is the identification of the device types within the smart home. This research hypothesizes that machine learning algorithms can be used to accurately perform the classification of smart home devices. A smart home environment was built using various BLE and Wi-Fi devices to create realistic traffic for machine learning classification. A device classification pipeline was designed to collect traffic and extract features. K-nearest neighbors, linear discriminant analysis, and random forests classifiers were built and tuned for experimental testing. Performance was evaluated using the Matthews correlation coefficient, mean recall, and mean precision metrics. Experimental results provide support towards the hypothesis that machine learning can classify device types to a high level of performance, but more work is necessary to build a more robust classifier.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 21, 2019
- Accession Number
- AD1074013
Entities
People
- Angelito Jr E. Aragon
Organizations
- Air Force Institute of Technology