Review: machine learning techniques applied to cybersecurity

By J. Martínez Torres, Carla Iglesias Comesaña, P. J. García–Nieto. 2019

192 citations recorded by the source corpus.

Machine learning techniques are a set of mathematical models to solve high non-linearity problems of different topics: prediction, classification, data association, data conceptualization. In this work, the authors review the applications of machine learning techniques in the field of cybersecurity describing before the different classifications of the models based on (1) their structure, network-based or not, (2) their learning process, supervised or unsupervised and (3) their complexity. All the capabilities of machine learning techniques are to be regarded, but authors focus on prediction and classification, highlighting the possibilities of improving the models in order to minimize the error rates in the applications developed and available in the literature. This work presents the importance of different error criteria as the confusion matrix or mean absolute error in classification problems, and relative error in regression problems. Furthermore, special attention is paid to the application of the models in this review work. There are a wide variety of possibilities, applying these models to intrusion detection, or to detection and classification of attacks, to name a few. However, other important and innovative applications in the field of cybersecurity are presented. This work should serve as a guide for new researchers and those who want to immerse themselves in the field of machine learning techniques within cybersecurity.

Subfields: Computer Networks and Communications, Information Systems, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Information Systems

In Signal Processing

Citations in the corpus

Mathematical optimization in classification and regression trees (2021, background): “…learning performance, are computationally cheap, and routines and packages to train them are available in popular languages such as Python and R, and are also appealing in terms of interpretability (Freitas 2014; Hu et al. 2019; Lin et al. 2020; Meinshausen 2010) because of their rule-based nature.”
A Survey of Deep Learning Methods for Cyber Security (2019, background): “Related surveys, such as those in [1–6] have described machine learning applications to cyber problems but did not include DL methods.”
A Survey on Machine Learning Techniques for Cyber Security in the Last Decade (2020, background): “Phishing and malware are considered as the most critical attacks [88].”