- Understand continuous variable probabilistic methods for identifying outliers
- Work with discrete variable probabilistic methods using RESTful and time series methods.
- Use entropy as a quick way to identify fields that may have anomalies.
- Use statistical properties to flag the data points that deviate from the expected.
- Using density based clustering, like DBSCAN, we can design the model such that the data points that do not fall into a cluster are the anomalies.*
- Continuous Probabilistic Methods
- Discrete Probabilistic Methods
- Detect Timeseries Anomalies
- Detect Through Clustering*
* need more practice