Automatic data categorization by Multi label Classification Using Semi-Supervised Singular Value Decomposition

In various domains including automatic multimedia data categorization, multilabel problems are arised, and have generated significant interest in computer vision and machine learning. However, existing methods do not adequately address two key challenges i.e. exploiting correlations between labels and making up for the lack of labelled data or even missing labelled data. We propose use of a semi supervised singular value decomposition to handle these two challenges. The proposed model takes advantage of the nuclear norm regularization on the Singular Value Decomposition electively capture the label correlations. Proposed method can exploit the label correlations and obtain promising and better label prediction results than the state-of-the-art methods.