Publications
Article · IEEE Signal Processing Letters · May 22, 2019
Asymptotically Bias-Corrected Regularized Linear Discriminant Analysis for Cost-Sensitive Binary Classification
Amin Zollanvari, Muratkhan Abdirash, Aresh Dadlani, Berdakh Abibullaev
Abstract
In this letter, the theory of random matrices of increasing dimension is used to construct a form of regularized linear discriminant analysis (RLDA) that asymptotically yields the lowest overall risk with respect to the bias of the discriminant in cost-sensitive classification of two multivariate Gaussian distributions. Numerical experiments using both synthetic and real data show that even in finite-sample settings, the proposed classifier can uniformly outperform RLDA in terms of achieving a lower risk as a function of regularization parameter and misclassification costs.
Statistical Methods and ApplicationsStatistical Methods and InferenceAdvanced Statistical Methods and Modelsregularized linear discriminant analysiscost-sensitive classificationbinary classification
More by Aresh Dadlani
- Joint Freshness and Age-Dispersion Control over Finite-State Markov Wireless ChannelsarXiv · 2026
- Dynamic Link Prediction with Temporally Enhanced Signed Graph Neural NetworksarXiv · 2026
- DeepIrrigo: Deep Reinforcement Learning for Continuous and Discrete Action Irrigation Optimization--A Case Study in IranSSRN Electronic Journal · 2026
- Age Performance Analysis in Resource-Constrained Adversarial IoT SystemsarXiv · 2025