“…Based on phasor measurement unit measurements, data‐driven approaches have been developed for DSA based on some machine learning algorithms. Some examples are decision trees [8–11], support vector machine [12], artificial neural networks [13], extreme learning machines (ELM) [5], random vector functional link neural network (RVFL) [14, 15], generative adversarial network [16], Lyapunov exponent [17], shapelet classification [18], imbalanced learning technique [19], probabilistic reliability method [20], integrated online learning method [21], random forest [22], semi‐supervised learning techniques [23], transfer learning algorithm [24], deep spatial‐temporal data‐driven approach [25], and online sequential learning method [26], which can carry out DSA control actions on time. As an emerging learning technology, ensemble learning also shows higher DSA performance [6].…”