Outlier Detection
Domain: Data & Datasets
🎯 Motivation
Identifying data points that are very different from the rest. Can be done statistically or with ML.
📋 Concrete Example
Detection: z-score > 3 = outlier. One article has 10,000 words (average: 500). z-score: 8.5. Outlier! Flag it for review or removal.
🔗 Analogy
A teacher noticing one student scored 0 while others scored 80-100. The 0 is an outlier. Outlier detection finds these anomalies automatically.