Anomaly Detection using Local Outlier Factor

Identify local outliers based on the densities of the neighborhood.

Here you will learn what is considered an outlier, why outlier detection is important, applications of outlier detection, different techniques to classify outliers, the difference between global, contextual, local, and collective outliers. What is the Local Outlier Factor(LOF), how it works, and implementation in python? Finally, the difference between Outlier detection and Novelty detection.

What is an Outlier?

An Outlier, also referred to as Anomaly, is an observation dissimilar or unusually different with respect to the rest of the observations in a dataset. Any data pattern that does not conform to the well-defined normal behavior.

Image by author

As per Hawkins

An outlier is an observation that deviates so much from other observations as to arouse suspicion that it was generated by a different mechanism

Based on Barnett and Lewis

An outlier is an observation (or subset of observations) which appears to be inconsistent with the remainder of that set of…