**2.1 Distance-based methods**

Distance-based method is based on the concept of the neighborhood of a sample and it was introduced by Knorr and Ng (Knorr & Ng, 1999). They gave the following definition: "*An object O in a dataset T is a DB(p,D)-outlier if at least fraction p of the objects in T lie at a distance greater than D from O*". The parameter p represents the minimum fraction of samples that is out of an outlier's D-neighborhood. This definition needs to fix a parameter and do not provide a degree of outlierness. Ramaswamy et al. (Ramaswamy et al., 2000) modified the definition of outlier: "*Outliers are the top n data points whose distance to the kth nearest neighbor is greatest*". Jimenez-Marquez et al. (Jimenez-Marquez et al., 2002) introduced the Mahalanobis Outlier Analysis (MOA) which uses Mahalnobis distance (Mahalanobis, 1936) as outlying degree of each point. Another outlier detection method based on Mahalanobis distance was proposed by Matsumoto et al. (Matsumoto et al., 2007). Mahalnobis distance is defined as the distance between each point and the center of mass. This approach considers outliers data points that are far away from their center of mass.
