Regularization algorithm of foundation settlement prediction model

I11-conditioning of least squares will lead to parameter solving failure. Based on analysis of foundation settlement prediction model and matrix inversion theory, using regularization theory, a new regularization unbiased estimation settlement prediction model algorithm is proposed. Unbiased and minimal variance of this new algorithm is described according to statistical theory. Existence of regularization parameter in the new algorithm is proofed in some condition. Calculation formula of regularization parameter is also given. Analyses with literatures and engineering examples show that the algorithm proposed not only reduces the matrix condition number and alleviates matrix ill-conditioning degree, but also can get the prediction model parameters effectively.