Unsupervised Outlier Model Selection (UOMS)

Background: Given an unsupervised outlier detection task, how should one select i) a detection algorithm, and ii) associated hyperparameter values (jointly called a model)? Effective outlier model selection is essential as different algorithms may work well for varying detection tasks, and moreover their performance can be quite sensitive to the values of the hyperparameters (HPs). Resources and environment for unsupervised outlier model selection (UOMS)

As a central place for UOMS related resources, you could find specific methods below:

• MetaOD (NeurIPS 2021): https://github.com/yzhao062/metaod
• ELECT (ICDM 2022): https://github.com/yzhao062/ELECT
• (Under review) A Review and Evaluation of Internal Evaluation Strategies (see Internal_Evaluation folder)