"Input to EIG must not contain NaN or Inf" error in fitrm for 18 subjects and 6 levels

Illustration
mkjep - 2022-07-16T11:57:22+00:00
Question: "Input to EIG must not contain NaN or Inf" error in fitrm for 18 subjects and 6 levels

Hi,   I am getting the following Matlab error: "Input to EIG must not contain NaN or Inf." when I run the fitrm function.   I am trying to test the differences between the 6 levels.   I have 18 subjects and 6 levels.   My table with the formatted data looks like this:   The code: subj = {'Subj1','Subj2','Subj3','Subj4','Subj5','Subj6','Subj7','Subj8','Subj9','Subj10','Subj11','Subj12','Subj13','Subj14','Subj15','Subj16','Subj17','Subj18'}'; t_loud_fc = table(subj,fc_y_250_mean,fc_y_500_mean,fc_y_1000_mean,fc_y_2000_mean,fc_y_3000_mean,fc_y_4000_mean,... 'VariableNames',{'subj','Cond1','Cond2','Cond3','Cond4','Cond5','Cond6'}); cond_loud_fc = dataset([1 2 3 4 5 6]','VarNames',{'Conditions'}); rm_loud_fc = fitrm(t_loud_fc,'Cond1-Cond6~subj','WithinDesign',cond_loud_fc); ranovatbl_loud_fc = ranova(rm_loud_fc); The output from the console window is:   Error using eig Input to EIG must not contain NaN or Inf. Error in RepeatedMeasuresModel>mauchlyTest (line 1921) lam = eig(C'*S*C); Error in RepeatedMeasuresModel.fit (line 1369) [this.Mauchly,this.Epsilon] = mauchlyTest(this.Cov,size(Xmat,1),rank(Xmat)); Error in fitrm (line 67) s = RepeatedMeasuresModel.fit(ds,model,varargin{:}); Error in RunStatAnalysis (line 51) rm_loud_fc = fitrm(t_loud_fc,'Cond1-Cond6~subj','WithinDesign',cond_loud_fc); I have traced the problem to line 1347 in RepeatedMeasuresModel.m. Here, the variable dfe is equal to zero and thus the covariance becomes 0 as well.. How do I avoid this?   The variable dfe is given by dfe = size(Xmat,1)-size(Xmat,2); , where Xmat is a 18x18 matrix and thus dfe becomes zero....

Related Questions

  • Retain dummy variable labels from converting categorical to dummyvar
  • "Input to EIG must not contain NaN or Inf" error in fitrm for 18 subjects and 6 levels
  • Avoiding for loop with anova1

    • Expert Answer

    Profile picture of John Williams John Williams answered . 2025-11-20

    Hello, I have a similar problem and don't understand how to solve it. I tried looking for missing values but don't seem to find them. This is my code:
     
    V1 = (41 + 47 + 42)/x;
    V2 = (54 + 97 + 58)/x;
    V3 = (46 + 49 + 26)/x;
    V4 = (28 + 40 + 45)/x;
    V5 = (104 + 126 + 70)/x;
    S1 = [V1 31 51 35 34];
    S2 = [V2 22 75 42 55];
    S3 = [V3 45 83 31 24];
    S4 = [V4 54 67 10 55];
    S5 = [V5 71 156 61 91];
    vehicle = [S1(1) S2(1) S3(1) S4(1) S5(1)];
    cloz_0_08 = [S1(2) S2(2) S3(2) S4(2) S5(2)];
    cloz_0_1 = [S1(3) S2(3) S3(3) S4(3) S5(3)];
    cloz_0_3 = [S1(4) S2(4) S3(4) S4(4) S5(4)];
    cloz_1 = [S1(5) S2(5) S3(5) S4(5) S5(5)];
    animals{1} = 'rat4';
    animals{2} = 'rat15';
    animals{3} = 'rat20';
    animals{4} = 'rat21';
    animals{5} = 'rat29';
    animals = animals'
    all = [vehicle' cloz_0_08' cloz_0_1' cloz_0_3' cloz_1'];
    t = table(animals,all(:,1),all(:,2), all(:,3), all(:,4), all(:,5),... 'VariableNames',{'animals','veh','cloz_0_08', 'cloz_0_1', 'cloz_0_3', 'cloz_1'});
    All = dataset([1 2 3 4 5]','VarNames',{'Treatments'});
    rm = fitrm(t,'veh-cloz_1~animals','WithinDesign',All)
    and when I run the last bit I get
    >> rm = fitrm(t,'veh-cloz_1~animals','WithinDesign',All) Error using eig Input to EIG must not contain NaN or Inf.
    Error in RepeatedMeasuresModel>mauchlyTest (line 1921) lam = eig(C'*S*C);
    Error in RepeatedMeasuresModel.fit (line 1369) [this.Mauchly,this.Epsilon] = mauchlyTest(this.Cov,size(Xmat,1),rank(Xmat));
    Error in fitrm (line 67) s = RepeatedMeasuresModel.fit(ds,model,varargin{:});


    Not satisfied with the answer ?? ASK NOW

    Get a Free Consultation or a Sample Assignment Review!

    Upload Assignment Talk to Expert