“…PCA is one of the most popular techniques for detecting outliers in various applications such as industrial processes (Li et al, 2000), environmental sensors (Harkatet al, 2006;Harrou et al,2013), distributed sensor networks (Chatzigiannakis and Papavassiliou, 2007), and high dimensional data (Ding and Kolaczyk, 2013). Most PCA-based models for outlier detection operate in batch mode (Chatzigiannakis and Papavassiliou, 2007;Harrou et al, 2013;Harkat et al, 2006), where the model is first trained using training data and is then used to test the remaining data for outliers.…”