LargeRegression: Large Regressions
Uses gradient descent to minimize the sum of squared
residuals for the regression problem. Can include an L2
penalty on the coefficient matrix. This function is very useful
when there is an initial guess of what B should be in Y = XB.
In general, this function performs faster than R's lm function.
GPU acceleration can be used to make this function extremely
fast. This package suggests cudaMatrixOps, which is not on
CRAN but can be downloaded at stanford.edu/~jeffwong.
|| <jeff.ct.wong at stanford.edu>|