topic Interpreting canonical structure and scores in discriminant analysis in Discussions https://community.jmp.com/t5/Discussions/Interpreting-canonical-structure-and-scores-in-discriminant/m-p/14969#M13895 <HTML><HEAD></HEAD><BODY><P>I'm trying to interpret the relative contributions of my variables in a discriminant function analysis of chemical data for three groups of rocks.&nbsp; In a stepwise, linear, common covariance discriminant function analysis I'm able to generate non-overlapping 95% confidence ellipses for my three groups with no incorrectly assigned data points using 6 variables.&nbsp; My understanding is that the scoring coefficients are indicative of the partial/unique contribution of each covariate to the canonical function.&nbsp; The total canonical structure (aka the structure coefficients) indicates the correlation between each covariate and the canonical function.&nbsp; Is that correct and if so, how would it apply to the canonical details below?</P><P></P><P>Total Canonical Structure</P><TABLE><TBODY><TR><TD>Row</TD><TD>SiO2</TD><TD>Sc</TD><TD>Hf</TD><TD>Ta</TD><TD>Th</TD><TD>U</TD></TR><TR><TD>Canon1</TD><TD>0.3516128771</TD><TD>0.0567257796</TD><TD>-0.347418256</TD><TD>-0.374134587</TD><TD>0.1334189492</TD><TD>0.3405553379</TD></TR><TR><TD>Canon2</TD><TD>0.1234474262</TD><TD>-0.003771588</TD><TD>0.3788469808</TD><TD>-0.033682518</TD><TD>0.3520133092</TD><TD>0.5659082601</TD></TR></TBODY></TABLE><P></P><P>Scoring Coefficients</P><TABLE><TBODY><TR><TD>Row</TD><TD>SiO2</TD><TD>Sc</TD><TD>Hf</TD><TD>Ta</TD><TD>Th</TD><TD>U</TD></TR><TR><TD>Canon1</TD><TD>0.0791912977</TD><TD>2.596079501</TD><TD>-0.66443715</TD><TD>-0.322485656</TD><TD>0.3434853562</TD><TD>0.309416116</TD></TR><TR><TD>Canon2</TD><TD>-0.017020659</TD><TD>-0.903204331</TD><TD>0.5762269008</TD><TD>-0.7698247</TD><TD>-0.064258925</TD><TD>0.530016795</TD></TR></TBODY></TABLE><P></P><P>Standardized Scoring Coefficients</P><TABLE><TBODY><TR><TD>Row</TD><TD>SiO2</TD><TD>Sc</TD><TD>Hf</TD><TD>Ta</TD><TD>Th</TD><TD>U</TD></TR><TR><TD>Canon1</TD><TD>0.7441814418</TD><TD>1.4836605646</TD><TD>-2.773050156</TD><TD>-0.828985708</TD><TD>2.0931281291</TD><TD>0.4034952175</TD></TR><TR><TD>Canon2</TD><TD>-0.159947609</TD><TD>-0.516181668</TD><TD>2.4049017984</TD><TD>-1.978921117</TD><TD>-0.391580486</TD><TD>0.6911703397</TD></TR></TBODY></TABLE></BODY></HTML> Tue, 20 Oct 2015 22:25:04 GMT amzipkin 2015-10-20T22:25:04Z Interpreting canonical structure and scores in discriminant analysis https://community.jmp.com/t5/Discussions/Interpreting-canonical-structure-and-scores-in-discriminant/m-p/14969#M13895 <HTML><HEAD></HEAD><BODY><P>I'm trying to interpret the relative contributions of my variables in a discriminant function analysis of chemical data for three groups of rocks.&nbsp; In a stepwise, linear, common covariance discriminant function analysis I'm able to generate non-overlapping 95% confidence ellipses for my three groups with no incorrectly assigned data points using 6 variables.&nbsp; My understanding is that the scoring coefficients are indicative of the partial/unique contribution of each covariate to the canonical function.&nbsp; The total canonical structure (aka the structure coefficients) indicates the correlation between each covariate and the canonical function.&nbsp; Is that correct and if so, how would it apply to the canonical details below?</P><P></P><P>Total Canonical Structure</P><TABLE><TBODY><TR><TD>Row</TD><TD>SiO2</TD><TD>Sc</TD><TD>Hf</TD><TD>Ta</TD><TD>Th</TD><TD>U</TD></TR><TR><TD>Canon1</TD><TD>0.3516128771</TD><TD>0.0567257796</TD><TD>-0.347418256</TD><TD>-0.374134587</TD><TD>0.1334189492</TD><TD>0.3405553379</TD></TR><TR><TD>Canon2</TD><TD>0.1234474262</TD><TD>-0.003771588</TD><TD>0.3788469808</TD><TD>-0.033682518</TD><TD>0.3520133092</TD><TD>0.5659082601</TD></TR></TBODY></TABLE><P></P><P>Scoring Coefficients</P><TABLE><TBODY><TR><TD>Row</TD><TD>SiO2</TD><TD>Sc</TD><TD>Hf</TD><TD>Ta</TD><TD>Th</TD><TD>U</TD></TR><TR><TD>Canon1</TD><TD>0.0791912977</TD><TD>2.596079501</TD><TD>-0.66443715</TD><TD>-0.322485656</TD><TD>0.3434853562</TD><TD>0.309416116</TD></TR><TR><TD>Canon2</TD><TD>-0.017020659</TD><TD>-0.903204331</TD><TD>0.5762269008</TD><TD>-0.7698247</TD><TD>-0.064258925</TD><TD>0.530016795</TD></TR></TBODY></TABLE><P></P><P>Standardized Scoring Coefficients</P><TABLE><TBODY><TR><TD>Row</TD><TD>SiO2</TD><TD>Sc</TD><TD>Hf</TD><TD>Ta</TD><TD>Th</TD><TD>U</TD></TR><TR><TD>Canon1</TD><TD>0.7441814418</TD><TD>1.4836605646</TD><TD>-2.773050156</TD><TD>-0.828985708</TD><TD>2.0931281291</TD><TD>0.4034952175</TD></TR><TR><TD>Canon2</TD><TD>-0.159947609</TD><TD>-0.516181668</TD><TD>2.4049017984</TD><TD>-1.978921117</TD><TD>-0.391580486</TD><TD>0.6911703397</TD></TR></TBODY></TABLE></BODY></HTML> Tue, 20 Oct 2015 22:25:04 GMT https://community.jmp.com/t5/Discussions/Interpreting-canonical-structure-and-scores-in-discriminant/m-p/14969#M13895 amzipkin 2015-10-20T22:25:04Z