Security for Multihop Wireless Networks

278  ◾  Security for Multihop Wireless Networks

11.2.1 Machine Learning-Based Anomaly Detection

Machine learning algorithms allow machines to learn from data. is learning

process extracts useful patterns from data and makes predictions on new data based

on previously obtained knowledge [9]. e machine learning-based techniques are

mainly used for training of anomaly-detection schemes. ree classical approaches

are used in training for anomaly-detection systems: supervised, semisupervised,

and unsupervised [2]. In this subsection, we critically review a few machine learn-

ing-based anomaly-detection schemes for WSNs.

11.2.1.1 Supervised Learning ...