Random Matrix Improved Covariance Estimation for a Large Class of Metrics

Authors: Malik Tiomoko, Romain Couillet, Florent Bouchard, Guillaume Ginolhac

ICML 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Section 4 provides experimental validations and applications, including an improved version of LDA/QDA based on the proposed enhanced estimates. sections 4.1. Validation on synthetic data and 4.2. Application to LDA/QDA contain figures and tables demonstrating empirical results.
Researcher Affiliation Academia 1Centrale Sup elec, University Paris Saclay, France 2GIPSAlab, University Grenoble-Alpes, France 3LISTIC, University Savoie Mont-Blanc, France.
Pseudocode Yes Algorithm 1 Proposed estimation algorithm. and Algorithm 2 Improved estimation from linear shrinkage.
Open Source Code Yes Matlab codes for the proposed estimation algorithms are available at https://github. com/maliktiomoko/RMTCov Est and are based on Manopt, a Matlab toolbox for optimization on manifolds (Boumal et al., 2014).
Open Datasets Yes The bottom right displays are applications to EEG data from (Andrzejak et al., 2001).
Dataset Splits No The paper mentions 'training vectors' for LDA/QDA but does not explicitly provide specific train/validation/test dataset splits or cross-validation details for reproduction.
Hardware Specification No The paper does not specify the hardware (e.g., CPU, GPU models, or memory) used for running the experiments.
Software Dependencies Yes Matlab codes for the proposed estimation algorithms are available at https://github. com/maliktiomoko/RMTCov Est and are based on Manopt, a Matlab toolbox for optimization on manifolds (Boumal et al., 2014).
Experiment Setup No The paper describes the gradient descent algorithm, step size 't' (fixed or optimized by backtracking line search), and initializations (M0 = Ip or linear shrinkage), but it does not provide specific numerical values for hyperparameters like fixed step size or details for the backtracking line search.