mlpack Multioutput Linear Regression

Multioutput Linear Regression

Open Z-Jianxin opened this issue 3 years ago • 1 comments

Add Multioutput Linear Regression functionality and relevant test cases.

Jul 31 '22 01:07 Z-Jianxin

Templatized Linear Regression so now the input data matrix for linear regression does not need to be arma::mat. The data matrix can now be arma::SpMat<float>, arma::SpMat<double>, arma::Mat<float>, or arma::Mat<double>.
The LinearRegression is renamed as LinearRegressionModel and the name LinearRegression now binds to the realized template LinearRegressionModel<arma::Mat, double>. So previous code and test cases do not break.
Add functionality to MultiOutput Regression. Given $A\in \mathbb{R}^{n\times k}$ , previously the linear regression model only supports $\min_{x \in \mathbb{R}^k }||Ax - y||_2^2$ with $y\in \mathbb{R}^n$ . Now we can do $\min_{X \in \mathbb{R} ^{k \times m} }||AX - Y||_2^2$ with $Y\in \mathbb{R}^{n\times m}$ .

Aug 04 '22 12:08 Z-Jianxin

This issue has been automatically marked as stale because it has not had any recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions! :+1:

Sep 03 '22 12:09 mlpack-bot[bot]

Sorry for the slow response on this one. I appreciate the contribution! But, is this any different than if the user simply made one LinearRegression model for each column of Y?

Sep 14 '22 21:09 rcurtin

mlpack mlpack copied to clipboard

Multioutput Linear Regression

mlpack
mlpack copied to clipboard