There a few (actually many) techniques for anomaly detection. Lumped into this are also novelty and outlier detection.
- Fraud Detection
- Loan Application Processing
- Intrusion Detection
- Activity Monitoring
- Network Performance
- Fault Diagnosis
- Novelties in Images
- Novelty in Text (and topics)
- Misslabeled data in training data
- Time series monitoring
- Medical Condition Monitoring
- Satellite Image Analysis
- etc.
- etc.
- Histogram
- Box plot
- Grubb's test
- Ordinary Least Squares
- Moving Average (+ STD)
- Poisson (or distribution based)
- Random Forest
- One class SVM
- Any other classifier really
- Affinity Propagation
- DBSCAN
- K-means
- Adapted kNN
Use an ensemble of anomaly detection techniques!
- Anomaly Detection: A survey
- A Survey of Outlier Detection Methodologiess
- Trend and Event Detection in Social Streams