Linear least squares

Linear least squares is a statistical method used to find the best-fitting linear relationship between a dependent variable and one or more independent variables. The goal of linear least squares is to minimize the sum of the squares of the differences (residuals) between the observed values and the values predicted by the linear model.