We want to justify why Kevin's QR method is better collect some timing data comparing it to other methods, including how well it scales with the number of rows in the least-squares problem.
Comparison candidates:
SuiteSparse QR
Recursive least squares (the rank-1 update thing)
Iterative least squares (show that these methods break or don't give accurate enough answers)
We want to justify why Kevin's QR method is better collect some timing data comparing it to other methods, including how well it scales with the number of rows in the least-squares problem.
Comparison candidates: