Outlier Detection with Non-Parametric Statistical Method

These methods dont make as many assumptions about the dat as Outlier Detection with Parametric Statistical Method and thus are applicable in more scenarios.

Use a Histogram to check if objects are Outliers. A problem with this method is choosing the correct bin size for the histogram. If the bin size is too small some normal objects might be treated as outliers (False Positive) and if the bin size is too big some outliers might fall in a frequent bin (False Negative).

Alternative to histogram is the Kernel Density Estimation.