针对路基沉降分析和预测中的正则回归模型算法,分析了正则参数的选取方法。根据矩阵条件数理论和最小二乘原理,提出了以控制信息矩阵条件数和残差平方和大小的正则参数选取新策略。结合二元回归模型的计算结果表明:提出的正则参数选取策略不仅降低了矩阵条件数,也能获得较小的残差平方和值。 Aiming at the regularization model algorithm in subgrade settlement analysis and prediction, the selection method of regular parameters is analyzed. According to the matrix condition number theory and the least square principle, a new strategy is proposed based on the parameters of the control information matrix and the square sum of residuals. Combined with the binary regression model, the results show that the proposed strategy can not only reduce the number of matrix conditions but also obtain a smaller residual sum of squares.