A Feature-Ranking Framework for IoT Device Classification

8 Pages Posted: 3 Dec 2018

See all articles by Bharat Atul Desai

Bharat Atul Desai

Singapore University of Technology and Design (SUTD)

Dinil Mon Divakaran

Singapore Telecommunications Limited (Singtel)

Ido Nevat

Heriot-Watt University - Department of Actuarial Mathematics and Statistics

Gareth Peters

Department of Actuarial Mathematics and Statistics, Heriot-Watt University; University College London - Department of Statistical Science; University of Oxford - Oxford-Man Institute of Quantitative Finance; London School of Economics & Political Science (LSE) - Systemic Risk Centre; University of New South Wales (UNSW) - Faculty of Science

Mohan Gurusamy

National University of Singapore (NUS)

Date Written: November 8, 2018

Abstract

IoT market is rapidly changing the cyber threat landscape. The challenges to security and privacy arise not only because IoT devices are large in number, but also because IoT devices are heterogeneous in type and functionality. Machine learning algorithms are attractive methods to solve various problems such as device identification, anomaly detection, and attack detection. Often, all available features extracted from network traffic are fed as input to train the models, which in practice is not regarded as the best approach. Associated with features are different kinds of cost, such as costs for obtaining the data, extracting and storing features, compute resources to run a model with high dimensional features, etc. Instead, if a smaller set of features could achieve performance close to that obtained with all features, that might help to reduce cost as well as to make better interpretation of results. In this work, we address the problem of selecting features extracted from IoT network traffic, based on the utility of a feature in achieving the goal of the machine learning models. We develop a unifying framework of fundamental statistical tests for ranking features. We specifically consider the use case of IoT device classification, and demonstrate the effectiveness of our framework by evaluating it using different classifiers on traffic obtained from real IoT devices.

Keywords: Classification, Feature Selection, Cyber Risk, Internet of Things (IoT)

JEL Classification: C10

Suggested Citation

Desai, Bharat Atul and Divakaran, Dinil Mon and Nevat, Ido and Peters, Gareth and Gurusamy, Mohan, A Feature-Ranking Framework for IoT Device Classification (November 8, 2018). Available at SSRN: https://ssrn.com/abstract=3281021 or http://dx.doi.org/10.2139/ssrn.3281021

Bharat Atul Desai

Singapore University of Technology and Design (SUTD) ( email )

20 Dover Drive
Singapore, 138682
Singapore

Dinil Mon Divakaran

Singapore Telecommunications Limited (Singtel) ( email )

Singapore

Ido Nevat

Heriot-Watt University - Department of Actuarial Mathematics and Statistics ( email )

Edinburgh, Scotland EH14 4AS
United Kingdom

Gareth Peters (Contact Author)

Department of Actuarial Mathematics and Statistics, Heriot-Watt University ( email )

Edinburgh Campus
Edinburgh, EH14 4AS
United Kingdom

HOME PAGE: http://garethpeters78.wixsite.com/garethwpeters

University College London - Department of Statistical Science ( email )

1-19 Torrington Place
London, WC1 7HB
United Kingdom

University of Oxford - Oxford-Man Institute of Quantitative Finance ( email )

University of Oxford Eagle House
Walton Well Road
Oxford, OX2 6ED
United Kingdom

London School of Economics & Political Science (LSE) - Systemic Risk Centre ( email )

Houghton St
London
United Kingdom

University of New South Wales (UNSW) - Faculty of Science ( email )

Australia

Mohan Gurusamy

National University of Singapore (NUS) ( email )

1E Kent Ridge Road
NUHS Tower Block Level 7
Singapore, 119228
Singapore

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
83
Abstract Views
470
rank
324,961
PlumX Metrics