Abstract
Discriminant analysis is an important and well-studied algorithm in pattern recognition area, and many linear discriminant analysis methods have been proposed over the last few decades. However, in the previous works, the between-scatter matrix is not updated when seeking the discriminant vectors, which causes redundancy for the well separated pairs. In this paper, a between-scatter matrix updating scheme is proposed based on the separable status of the obtained vectors. In our scheme, separable status determination of obtained vectors is decisive. Here, we notice that appropriate separation of a multi-dimensional feature (with homoscedastic Gaussian distribution) may help to find better discriminant vectors, and the separability of a multi-dimensional feature can be deduced from the separability of its elements. To make the discriminant vectors statistically uncorrelated, the algorithm is applied to the St-orthogonal space of the obtained vectors in an iterative way. We also extend our method to more general cases, like heteroscedastic distributions, by an appropriate kernel function. Experimental results on multiple databases demonstrate the effectiveness of the proposed method.
Original language | English |
---|---|
Pages (from-to) | 154-162 |
Journal | Neurocomputing |
Volume | 173 |
Issue number | Part 2 |
DOIs | |
Publication status | Published - 2 Jan 2016 |
Keywords
- LDA
- Classification
- Dimension reduction
- Updated between-scatter matrix