Novelty detection is the unsupervised problem of identifying anomalies in test data which significantly differ from the training set. While numerous novelty detection methods were designed to model continuous numerical data, tackling datasets composed of mixed-type features, such as numerical and categorical data, or temporal datasets describing discrete event sequences is a challenging task. In addition to the supported data types, the key criteria for efficient novelty detection methods are the ability to accurately dissociate novelties from nominal samples, the interpretability, the scalability and the robustness to anomalies located in the training data.
In this thesis, we investigate novel ways to tackle these issues. In particular, we propose (i) a survey of state-of-the-art novelty detection methods applied to mixed-type data, including extensive scalability, memory consumption and robustness tests (ii) a survey of state-of-the-art novelty detection methods suitable for sequence data (iii) a probabilistic nonparametric novelty detection method for mixed-type data based on Dirichlet process mixtures and exponential-family distributions and (iv) an autoencoder-based novelty detection model with encoder/decoder modelled as deep Gaussian processes. The learning of this last model is made tractable and scalable through the use of random feature approximations and stochastic variational inference. The method is suitable for large-scale novelty detection problems and data with mixed-type features. The experiments indicate that the proposed model achieves competitive results with state-of-the-art novelty detection methods.
