Matlab相关应用

用Matlab做回归计算

x = [1656,2122,2864,4033,6099]

y = [2112,2170,2291,2456,2759]

>> clear
>> x=[1656,2122,2864,4033,6099];
>> y=[2112,2170,2291,2456,2759];
>> [P,S]=polyfit(x,y,1)
%P为拟合回归系数即y=P(1)*x+P(2)
P =
  1.0e+003 *
    0.0001    1.8666
S =
        R: [2x2 double]
       df: 3
    normr: 9.4223

>> [Y,delta]=polyconf(P,x,S)
%给出回归Y的95%的置信区间为[Y-delta，Y+delta]
Y =
  1.0e+003 *

    2.1090    2.1772    2.2858    2.4569    2.7592
delta =

   20.6907   19.8923   19.1147   19.2502   23.2005
>> x1=1600:100:6100;
>> f=polyval(P,x1);
>> plot(x,y,'ro',x1,f,'-')%绘图查看拟合效果
>> hold on
>> plot(x,Y+delta,'*g')
>> plot(x,Y-delta,'*g')%给出拟合的置信区间