Understanding outliers
Many data points are far away from the group centroid or average, affecting the value of the standard value. Remember that the optimal case is to have a compact group with a small standard value. These data points are outliers that don't comply with the characteristics of the group or the general data. These outliers deserve further investigation because they could have suspicious behavior (for example, fraud or peaks in demand) that could affect the performance of the system, such as network traffic. Look at Figure 4.14 to examine the outliers of the groups of one variable and the outliers of the 3D chart:
Figure 4.14 – Outliers for one variable and three variables
Group Number 3 is the outlier of the one-variable classification. Notice that the points have a large separation from the centroid. The standard deviation has a large amplitude. This group has different behavior compared with the other two, which are compact...
