def left_least_squares(x,y,rcond=-1,fast=False):
'find the A that best fits y-A*x'
if fast:
return la.lstsq( np.matmul(x,x.T) ,np.matmul(x,y.T) ,rcond=rcond )[0].T # faster, but less stable
else:
return la.lstsq( x.T,y.T,rcond=rcond)[0].T
评论列表
文章目录