On the Equivalence of Linear Discriminant Analysis and Least Squares

Authors: Kibok Lee, Junmo Kim

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results demonstrate the equivalence of the LDA solution and the proposed LS solution.
Researcher Affiliation Collaboration Kibok Lee1,2 and Junmo Kim1 1Department of Electrical Engineering, KAIST, Daejeon, Korea 2Samsung Electronics DMC R&D Center, Suwon, Korea
Pseudocode Yes Algorithm 1 Fast approximation of regularized uncorrelated LDA (RULDA) (p = c 1) 1. Compute YB = ZBL, where ZB = {z Bij} in (5). 2. Solve W = argmin W XT CW YT B 2 F + γ W 2 F . 3. return W
Open Source Code No No. The paper does not provide any explicit statements about releasing source code or links to a code repository for the described methodology.
Open Datasets Yes We used three data sets for our experiment: the extended Yale Face Database B (Georghiades, Belhumeur, and Kriegman 2001), the MNIST database of handwritten digits (Le Cun and Cortes 1998), and Isolet (Bache and Lichman 2013).
Dataset Splits No No. The paper provides the number of training and test samples but does not specify a clear split methodology (e.g., percentages, random seed, k-fold cross-validation) or mention a validation set.
Hardware Specification Yes All experiments were done in MATLAB on a PC with an Intel Core i73610QM CPU at 2.30 GHz and with 8 GB RAM.
Software Dependencies No No. The paper mentions using 'MATLAB' but does not provide a specific version number for MATLAB or any other software libraries or tools used in the experiments.
Experiment Setup Yes The regularization parameter is set to be 10 4 on extended Yale B and 1 on the others.