A least squares method is a method of fitting a curve to data points so as to minimize the sum of the squares of the distances of the points to the curve.