The rate of return of style factor K and the residual/characteristic rate of return of stock N can be obtained by cross-sectional regression.In order to reduce the estimation error of the model, the weighted least squares (WLS) is usually used to estimate the regression coefficient.