Truncated Cauchy Non-negative Matrix Factorization
This addresses robust subspace learning for noisy datasets, but it is incremental as it modifies an existing method to handle outliers.
The paper tackles the problem of non-negative matrix factorization (NMF) being sensitive to outliers in corrupted data by proposing a Truncated CauchyNMF loss that truncates large errors, and it shows effectiveness in image clustering on noisy datasets with gross corruptions.
Non-negative matrix factorization (NMF) minimizes the Euclidean distance between the data matrix and its low rank approximation, and it fails when applied to corrupted data because the loss function is sensitive to outliers. In this paper, we propose a Truncated CauchyNMF loss that handle outliers by truncating large errors, and develop a Truncated CauchyNMF to robustly learn the subspace on noisy datasets contaminated by outliers. We theoretically analyze the robustness of Truncated CauchyNMF comparing with the competing models and theoretically prove that Truncated CauchyNMF has a generalization bound which converges at a rate of order $O(\sqrt{{\ln n}/{n}})$, where $n$ is the sample size. We evaluate Truncated CauchyNMF by image clustering on both simulated and real datasets. The experimental results on the datasets containing gross corruptions validate the effectiveness and robustness of Truncated CauchyNMF for learning robust subspaces.