Feature Extraction Using Discriminant Graph Laplacian Principal Component Analysis with Application to Biomedical Datasets

Abstract

Abstract In this paper, we propose a manifold learning method called discriminant graph Laplacian principal component analysis (DGLPCA) for feature extraction. The proposed method projects high dimensional data into a lower dimensional subspace while preserving much of the intrinsic structure of the data. Moreover, DGLPCA integrates maximum margin criterion into its objection function to improve class separability in the lower dimensional space. The effectiveness of the proposed method is demonstrated on two publicly available biomedical datasets taken from UCI machine learning repository. The results show that our proposed method provides more discriminative power compared to other similar approaches.