interpretable anomaly detection
subspace anomaly detection
2012.ICDM.Outlier Detection in Arbitrarily Oriented Subspaces(Correlation Outlier Probability (COP))[pdf]
- determine whether an object is anomalous on arbitrarily oriented subspaces spanned by its neighbors.
2013.CIKM.Flexible and adaptive subspace search for outlier analysis [pdf]
- utilize the differences of outlier scores in random subspaces to perform a combinatorial refinement of relevant subspaces.
interpretable anomaly detection
subspace
2013.CIKM.Flexible and adaptive subspace search for outlier analysis [pdf]
- find a subspace which maximizes differences in outlier score distributions of all points across subspaces
visualize
2019.ECML PKDD.Beyond outlier detection: lookout for pictorial explanation[pdf,code]
- use a set of focus-plots, each of which “blames” or “explains away” a subset of the input outliers.
others
1999.VLDB.Finding intensional knowledge of distance-based outliers[pdf]
- define the outlier categories C = {“trivial outlier,” “weak outlier,” “strongest outlier”} to help gain better insights about the nature of outliers.
2013.ECML/PKDD.Local outlier detection with interpretation[pdf]
2014.ICDE.Discriminative features for identifying and interpreting outliers[pdf]
2016.AAAI.A framework for outlier description using constraint programming[pdf]
surveys
2022.DKE.Anomaly explanation: A review[pdf]
- review anomaly detection methods and anomaly explanation method (explanation by
feature importance, explanation by feature values, explanation by data points
comparison and explanation by structure analysis)