topic Re: Erro with maximum likelihood analysis in SAS Programming

Erro with maximum likelihood analysis

rjinks — Mon, 16 Nov 2020 19:21:09 GMT

I have output values for my maximum likelihood analysis that are 1.00 and I have an error message that correlation matrix is singular. Where or what is the cause of these these 2 issues?

Datalines;
LC 0 0 0.024266667 0.015533333 0.02602 0.008633333 0.010333333 0.017233333 0 0.006366667 0.041333333 0.016933333 0.023333333 0.053833333
SK1 0.004966667 0.004966667 0.0078 0.005666667 0.008533333 0 0 0.0078 0 0 0.003566667 0.003566667 0.008593333 0.027133333 
WR1 0.0042 0.0077 0.04 0.0175 0.02539 0.0063 0.0091 0.0224 0 0 0 0.0091 0.018216667 0.093156667
WR2 0 0 0.00595 0.00595 0.011833333 0.0042 0.00345 0.005326667 0.002176667 0 0 0.003733333 0.006066667 0.013743333
PI 0 0 0 0 0 0 0 0 0 0 0 0 0.006003333 0
;
title3 'Maximum Likelihood Factor Analysis with One Factor';
proc factor data=SsamPAHan method=ml heywood n=1;
title3 'Maximum Likelihood Factor Analysis with Two Factor';
proc factor data=SsamPAHan method=ml heywood n=2;
title3 'Maximum Likelihood Factor Analysis with Three Factor';
proc factor data=SsamPAHan method=ml heywood n=3;

Re: Erro with maximum likelihood analysis

PaigeMiller — Mon, 16 Nov 2020 19:28:14 GMT

Can you show us the full complete log for this portion of code, with nothing chopped out? Please copy the log as text and paste it into the windows that appears when you click on the </> icon here. DO NOT SKIP THIS STEP.

Also, showing us the output would help.

Re: Erro with maximum likelihood analysis

Reeza — Mon, 16 Nov 2020 19:31:29 GMT

Your data structure seems weird, can you please explain your data as well.

Re: Erro with maximum likelihood analysis

rjinks — Mon, 16 Nov 2020 19:49:22 GMT

1 OPTIONS NONOTES NOSTIMER NOSOURCE NOSYNTAXCHECK;

73 Data SsamPAHan;

74 Input Site $ Ace Anthra Benzoaa Benzoapy

75 Benzobflu Benzoghi Benzokfluo Chry

76 Dibenzo Fluoran Fluorene Indeno123

77 Phenan Pyr;

78 Datalines;

NOTE: The data set WORK.SSAMPAHAN has 5 observations and 15 variables.

NOTE: DATA statement used (Total process time):

real time 0.00 seconds

cpu time 0.00 seconds

84 ;

85 title3 'Maximum Likelihood Factor Analysis with One Factor';

86 proc factor data=SsamPAHan method=ml heywood n=1;

87 title3 'Maximum Likelihood Factor Analysis with Two Factor';

WARNING: The number of observations is not greater than the number of variables.

ERROR: Correlation matrix is singular.

NOTE: Prior communality estimates will be 1.0.

ERROR: METHOD=ML requires a nonsingular correlation matrix.

NOTE: The SAS System stopped processing this step because of errors.

NOTE: PROCEDURE FACTOR used (Total process time):

real time 0.05 seconds

cpu time 0.05 seconds

88 proc factor data=SsamPAHan method=ml heywood n=2;

89 title3 'Maximum Likelihood Factor Analysis with Three Factor';

WARNING: The number of observations is not greater than the number of variables.

ERROR: Correlation matrix is singular.

NOTE: Prior communality estimates will be 1.0.

ERROR: METHOD=ML requires a nonsingular correlation matrix.

NOTE: The SAS System stopped processing this step because of errors.

NOTE: PROCEDURE FACTOR used (Total process time):

real time 0.03 seconds

cpu time 0.03 seconds

90 proc factor data=SsamPAHan method=ml heywood n=3;

97 OPTIONS NONOTES NOSTIMER NOSOURCE NOSYNTAXCHECK;

WARNING: The number of observations is not greater than the number of variables.

ERROR: Correlation matrix is singular.

ERROR: METHOD=ML requires a nonsingular correlation matrix.

109

Re: Erro with maximum likelihood analysis

PaigeMiller — Mon, 16 Nov 2020 21:05:06 GMT

NOTE: The data set WORK.SSAMPAHAN has 5 observations and 15 variables.
WARNING: The number of observations is not greater than the number of variables.
ERROR: Correlation matrix is singular.
NOTE: Prior communality estimates will be 1.0.
ERROR: METHOD=ML requires a nonsingular correlation matrix.

If you have 5 records and 15 variables, then mathematically, you are guaranteed to have a singular correlation matrix. So you can't do this analysis unless you have at least (I think) 15 records.

Re: Erro with maximum likelihood analysis

rjinks — Tue, 17 Nov 2020 12:05:58 GMT

Thank you. I am reading up on matrices to insure that going forward I have the right conditions for the expected output.