Computer Science
5 years ago

Unbiased estimator revolutionizes active learning for optimal covariance matrix estimation.

Active Covariance Estimation by Random Sub-Sampling of Variables
Eduardo Pavéz, Antonio Ortega

Paper Summary

Related papers

Theme background image
Computer Science
8 years ago

New method boosts accuracy in facial recognition technology.

Dimensionality Reduction
Neural Networks
Hyperspectral Analysis
2 views
Theme background image
Engineering
a year ago

New method for estimating noise covariance could improve data analysis accuracy.

System Identification
Health Monitoring
Least Squares
2 views
Theme background image
Mathematics
8 years ago

New method for estimating covariance matrix reduces sensitivity to outliers.

Multicollinearity
Least Squares
Multivariate Analysis
2 views

The article explores a method to estimate covariance matrices when only some parts of the data are observed. By using a special type of random variable, the researchers developed an unbiased estimator for the covariance matrix and found a way to control the error in the estimation. They applied this method to a scenario where only a few variables are observed compared to the total number, and came up with a way to choose the best variables to observe for accurate estimation.