Different approaches to the detection of anomalies.
The presentation was made within the Virtual Colloquium of the Mexican Statistical Association (Asociación Mexicana de Estadística, AME) 2020.
The main topic of the presentation was the algorithms for the detection of outliers, different approaches were shown for this detection, each approach had its respective examples, which were:
-
Statistical approach: Box-and-Whisker plot and MAD.
-
Density approach: LOF and Isolation Forest.
-
Approach through Machine Learning: Autoencoders and Self-Organizing Maps.