MATLAB绘制残差直方图

%% step 1
hospital = readtable('hospital.xls','ReadRowNames',true);
hospital.smoke = nominal(hospital.smoke,{'No','Yes'}); 
%% step 2
model='sys ~ 1 + age + wgt + smoke';
mdl = fitlm(hospital, model)
plotResiduals(mdl)

outlier = mdl.Residuals.Raw > 9;  %数据清洗条件
OL=find(outlier);
mdl = fitlm(hospital, model,'Exclude', OL);
mdl.ObservationInfo(OL,:)

mdl1 = step(mdl,'NSteps',10);

mdl1.Formula.disp
%% step 3
ages = [20;40;60];
smoker = {'Yes';'No';'Yes'};
predictnew = mdl1.feval(ages,smoker)

mdl1.Coefficients(:,1)

>> Case_52

mdl = 


线性回归模型:
    sys ~ 1 + age + wgt + smoke

估计系数:
                    Estimate         SE         tStat        pValue  
                   ___________    ________    _________    __________

    (Intercept)         116.28      3.6668       31.712    1.1478e-52
    age               0.085808    0.066986        1.281       0.20329
    wgt            -0.00098391    0.018594    -0.052915       0.95791
    smoke_Yes           9.8925      1.0353       9.5552    1.3434e-15


观测值数目: 100，误差自由度: 96
均方根误差: 4.78
R 方: 0.507，调整 R 方 0.492
F 统计量(常量模型): 33，p 值 = 9.91e-15

ans =

  2×4 table

               Weights    Excluded    Missing    Subset
               _______    ________    _______    ______

    WXM-486       1        true        false     false 
    XBJ-540       1        true        false     false 

1。正在删除 wgt, FStat = 0.25432, pValue = 0.61523
sys ~ 1 + age + smoke

predictnew =

  126.9535
  119.2712
  132.0364


ans =

  3×1 table

                   Estimate
                   ________

    (Intercept)     114.19 
    age            0.12707 
    smoke_Yes       10.224 

开发工具:MATLAB 2022b
微信截屏工具 Alt+A