Dictionary:ℓp fit

{{#category_index:L|ℓ_p fit}} The parameters that produce estimations y_i to a set of data points y_i such that the differences between the data and the estimates (the errors e_i=y_i–y_i) minimizes

$\Psi =\sum _{j}^{p}|w_{j}e_{j}|$

where w_i are weighting factors. If p=1, this yields the least-absolute deviation fit; for p=2, least-squares; for p=infinity, the minimax or Chebychev results.