Processing math: 100%
GOAL: Generalized Jointly Sparse Linear Discriminant Regression for Feature Extraction | IEEE Journals & Magazine | IEEE Xplore

GOAL: Generalized Jointly Sparse Linear Discriminant Regression for Feature Extraction


Impact Statement:Feature extraction is the basic procedure in artificial intelligence tasks. A proper feature extraction method is helpful for downstream work. The proposed GOAL is a robu...Show More

Abstract:

Ridge regression (RR)-based methods aim to obtain a low-dimensional subspace for feature extraction. However, the subspace's dimensionality does not exceed the number of ...Show More
Impact Statement:
Feature extraction is the basic procedure in artificial intelligence tasks. A proper feature extraction method is helpful for downstream work. The proposed GOAL is a robust regression method based on L_{2,1}-norm and capped-L_{2}-norm measurements that obtains a jointly sparse feature subspace to extract the discriminative features. Sparsity, locality, and discriminability are integrated into one model to learn a full-rank robust feature extractor.

Abstract:

Ridge regression (RR)-based methods aim to obtain a low-dimensional subspace for feature extraction. However, the subspace's dimensionality does not exceed the number of data categories, hence compromising its capability of feature representation. Moreover, these methods with L_{2}-norm metric and regularization cannot extract highly robust features from data with corruption. To address these problems, in this article, we propose generalized jointly sparse linear discriminant regression (GOAL), a novel regression method based on joint L_{2,1}-norm and capped-L_{2}-norm, which can integrate sparsity, locality, and discriminability into one model to learn a full-rank robust feature extractor. The sparsely selected discriminative features are robust enough to characterize the decision boundary between classes. Locality is related to manifold structure and Laplacian smoothing, which can enhance the robustness of the model. By using the multinorm metric and regularization regression framework, the proposed method obtains the projection with joint sparsity and guarantees that the rank of the projection matrix will not be limited by the number of classes. An iterative algorithm is proposed to compute the optimal solution. Complexity analysis and proofs of convergence are also given in the article. Experiments on well-known datasets demonstrate our model's superiority and generalization ability.
Published in: IEEE Transactions on Artificial Intelligence ( Volume: 5, Issue: 10, October 2024)
Page(s): 4959 - 4971
Date of Publication: 11 June 2024
Electronic ISSN: 2691-4581

Funding Agency:


References

References is not available for this document.