Hi,
I would like to use PCA to generate a composite indicator which would equal the weighted average of the scores predicted for each PC, where used weights are equivalent to the proportion of variance explained by each component in the PCA. I would normalize PC scores before calculating the weighted average.
Two questions:
1) What command would be most appropriate for this purpose? Are there any options for 'predict' which would allow me to do it straightforward?
2) I need to calculate this score from a number of PCAs run on a number of datasets, and therefore I do not know a priori the number of PCs to be retained according to a certain criterion which I will adopt, e.g. mineigen(1). Therefore I am not able, in 'predict' to define a priori and label the new variables generated by PC scores extraction.
Hope it is clear
Thank you very much for your help
I would like to use PCA to generate a composite indicator which would equal the weighted average of the scores predicted for each PC, where used weights are equivalent to the proportion of variance explained by each component in the PCA. I would normalize PC scores before calculating the weighted average.
Two questions:
1) What command would be most appropriate for this purpose? Are there any options for 'predict' which would allow me to do it straightforward?
2) I need to calculate this score from a number of PCAs run on a number of datasets, and therefore I do not know a priori the number of PCs to be retained according to a certain criterion which I will adopt, e.g. mineigen(1). Therefore I am not able, in 'predict' to define a priori and label the new variables generated by PC scores extraction.
Hope it is clear
Thank you very much for your help
