I have conducted factor analysis and have identified 13 latent variables from 65 original variables. I have generated the factor scores using the command "predict" in STATA.
I would now like to proceed with regression analysis to see the association of these factors with some other psychiatric variables.
Since the factor scores are continuous as opposed to the previous categorical variables, I am confused as to how to use them in regression or any other further analysis. Please guide.
OVER-FITTING PROBLEM. You have 65 original variables + 13 latent = 78 "variables." This is overfitted. In common layman's term, the researcher cannot explain the phenomenon so include so many into the model and hope for the best. You need a better approach. the first place to start is literature review. See what factors or variables have other researchers used to explain this issue? What other variables you can contribute.
AKAIKE INFORMATION may help in explaining overfitting problem. see https://en.wikipedia.org/wiki/Akaike_information_criterion
https://www.statisticssolutions.com/what-is-multiple-linear-regression/:
" Multiple linear regression is the most common form of linear regression analysis. As a predictive analysis, the multiple linear regression is used to explain the relationship between one continuous dependent variable and two or more independent variables. The independent variables can be continuous or categorical (dummy coded as appropriate). "
I wou can use stepwise or -even better- best subset regression, to look for a reduced model (with as few predictors as possible).
With STATA:
https://www.stata.com/manuals13/rstepwise.pdf
or attachment
may be "mutual_info_regression" can be solve your question. this 'mutual_info_regression' is appropriate if the target value is continuous data.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis