Unsupervised outlier detection using random subspace and subsampling ensembles of Dirichlet process mixtures

Probabilistic mixture models are recognized as effective tools for unsupervised outlier detection owing their interpretability and global characteristics. Among these, Dirichlet process mixture models stand strong alternative to conventional finite mixture models for both clustering and outlier detection tasks. finite mixture models, Dirichlet process mixtures are infinite mixture models that automatically determine number of mixture components based on the data. Despite their advantages, the adoption of Dirichlet mixture models for unsupervised outlier detection has been limited by challenges related to computational inefficiency and sensitivity to outliers in the construction of outlier detectors. Additionally, Dirichlet Gaussian mixtures struggle to effectively model non-Gaussian data with discrete or binary features. To these challenges, we propose a novel outlier detection method that utilizes ensembles of Dirichlet Gaussian mixtures. This unsupervised algorithm employs random subspace and subsampling ensembles ensure efficient computation and improve the robustness of the outlier detector. The ensemble approach improves the suitability of the proposed method for detecting outliers in non-Gaussian data. Furthermore, our method uses variational inference for Dirichlet process mixtures, which ensures both efficient and computation. Empirical analyses using benchmark datasets demonstrate that our method outperforms approaches in unsupervised outlier detection.