E-commerce Anomaly Detection: A Bayesian Semi-Supervised Tensor Decomposition Approach using Natural Gradients
This work addresses anomaly detection for e-commerce platforms, offering an incremental improvement by incorporating sparse labeled data into tensor decomposition.
The paper tackled anomaly detection in e-commerce seller-reviewer data by developing a Bayesian semi-supervised tensor decomposition method with Polya-Gamma augmentation and partial natural gradient learning, resulting in improved performance over unsupervised baselines.
Anomaly Detection has several important applications. In this paper, our focus is on detecting anomalies in seller-reviewer data using tensor decomposition. While tensor-decomposition is mostly unsupervised, we formulate Bayesian semi-supervised tensor decomposition to take advantage of sparse labeled data. In addition, we use Polya-Gamma data augmentation for the semi-supervised Bayesian tensor decomposition. Finally, we show that the Pólya-Gamma formulation simplifies calculation of the Fisher information matrix for partial natural gradient learning. Our experimental results show that our semi-supervised approach outperforms state of the art unsupervised baselines. And that the partial natural gradient learning outperforms stochastic gradient learning and Online-EM with sufficient statistics.