A critical role is played by Wireless Sensor Networks (WSNs) in modern communication systems, to monitor and control remote environments. In order to identify breaches in WSNs, many methods have been suggested. When it comes to defence in depth, however, mission-critical applications should always have an Intrusion Detection System (IDS) installed. The aim of this work is to compare different machine learning (ML) algorithms developed for WSN-oriented IDS. A comparative analysis of multiple ML techniques, including Logistic Regression (LR), Linear Discriminant Analysis (LDA), Naïve Bayes, Decision Trees (DT), Support Vector Machines (SVMs), Random Forests (RFs), K-Nearest Neighbours (K-NN) and Convolutional Neural Networks (CNNs), is conducted to evaluate their effectiveness in detecting various types of attacks including denial-of-service (DoS), spoofing and blackhole attacks. Performance metrics such as accuracy, precision and recall, are employed to assess the suitability of each algorithm for resource-constrained WSN environments.