Of course, there are a number of generally accepted machine learning based approaches to the process of anomaly detection. These currently include the following:

• Density-based anomaly detection: This approach is based on the KNN algorithm; the nearest set of data points are evaluated using a scoring method dependent on the type of the data (categorical or numerical).
• Clustering-based anomaly detection: One of the most desired concepts in the domain of unsupervised learning for anomaly detection is Clustering.
• Support vector machine based anomaly detection: This algorithm uses a training set to learn soft boundaries in order to cluster the normal data instances then, using the testing instance, it calibrates itself to locate the abnormalities that fall outside the learned region.