Prediction and pattern discovery models are valid as long as the observations on which these models are built resemble the current analyzed observations. Detecting if changes occur is important for adaptability, and identifying the precise moment of their occurrence is critical for control systems to enable timely reactions. The former relates to concept drift, whereas the latter pertains to the definition of change points. Change points are detailed in Sect. 4.1, while change point detectors are covered in Sect. 4.2. Concept drift is covered in Sect. 4.3, with the main drift detector methods detailed in Sect. 4.4. Advanced topics regarding change points and concept drift are covered in Sect. 4.5.