Currently, bolasso is implemented to estimate bootstrapped Lasso models in parallel. I want to extend this to the coef and predict methods as well. Although the performance gains are probably much, much smaller for both of these cases, it makes it more consistent and, in the case of extremely large data, could improve performance a fair amount.
Currently,
bolasso
is implemented to estimate bootstrapped Lasso models in parallel. I want to extend this to thecoef
andpredict
methods as well. Although the performance gains are probably much, much smaller for both of these cases, it makes it more consistent and, in the case of extremely large data, could improve performance a fair amount.