Anomaly Detection Techniques on relational data as Quality Evaluation of a dataset

Abstract

An outlier is a point that deviates significantly from the pattern that has been formed from the majority of the data points. The presence of outliers can exacerbate statistical results which leads to misrepresented relationships between different data and faulty conclusions based on them. This is an urgent issue in a data driven world and people in the data sector are following harsh and tedious procedures to deal with that. In the present article, an half automated tool for outlier detection returning a single score for a structured dataset is proposed with minimal human intervention. This tool can either be used in a python environment or directly in a python script.