Anomaly Detection using Scikit Learn

Inlier: A Data point that falls within the normal range defined by the majority of the data. it fits the expected pattern

outlier: A data point that deviates significantly from the main cluster

Anomaly Detection: Finding unusual patterns in data

1. Outlier Detection / Unsupervised anomaly detection

- Training data might contain some outliers already
- Goal is to identify these outliers based on how far they deviate from the majority of the data

2. Novelty Detection / Semi-supervised anomaly detection

- Training data does not have outliers
- Real time data/Production data has outliers, which needed to be found.