中文摘要：

为了提高船舶操纵性指数K、T的预报精度以便更好地评价船舶性能。采用多元非线性回归模型．利用SPSS软件对43艘商船的K、T指数实船试验结果进行了回归分析。研究了对K、T有显著影响的非线性因素，给出了两个最优回归多项式。以某杂货船为例进行了操纵性指数的预报．预报效果令人满意，K、T误差分别为4．4％和1．1％。与原有的线性模型相比。预报误差分别减小了16．4％和70．3％。该文提出的非线性模型在回归方程的显著性以及对K’、T’值的预报准确性上均较线性模型有了很大提高．对操船者在航海实践中评估船舶的操纵性能有一定的参考价值。

英文摘要：

In order to enhance the prediction accur - acy of ship＇s maneuverability indices K and T, a multivariate nonlinear regression analysis was applied based on 43 merchant ships＇ zig-zag test results using SPSS software. Two optimum regression equations were given after comparing the rion-linear factors which have remarkable influence on K and T indices. The prediction was carried out for a general cargo ship to compare the nonlinear model with the original linear model. The forecast effect is satisfactory, the error is 4.4% and 1.1% for K and T respectively, compared with the original linear model, the prediction error separately dropped 16.4% and 70.3%. This nonlinear model gets remarkable improvement in the regression equation significance as well as the indices prediction precision. It is of reference value for ship handlers to evaluate their ships＇ maneuv - erability in navigation practice.