Model Selection for Anomaly Detection
This work addresses a specific bottleneck in anomaly detection for domains like image processing and network intrusion detection, but it is incremental as it adapts existing methods rather than introducing a new paradigm.
The paper tackled the problem of kernel selection for one-class classification in anomaly detection, where standard methods like cross-validation are not directly applicable due to the absence of abnormal class data. It generalized several kernel selection methods from binary-class to one-class classification and performed extensive comparisons using synthetic and real-world data, though no concrete performance numbers were provided.
Anomaly detection based on one-class classification algorithms is broadly used in many applied domains like image processing (e.g. detection of whether a patient is "cancerous" or "healthy" from mammography image), network intrusion detection, etc. Performance of an anomaly detection algorithm crucially depends on a kernel, used to measure similarity in a feature space. The standard approaches (e.g. cross-validation) for kernel selection, used in two-class classification problems, can not be used directly due to the specific nature of a data (absence of a second, abnormal, class data). In this paper we generalize several kernel selection methods from binary-class case to the case of one-class classification and perform extensive comparison of these approaches using both synthetic and real-world data.