I know the basis of Principal components analysis (PCA). But I have a question regarding a situation when we derive many variables from two variables. For example, in the analysis of crop tolerance, when we have YP (yield in normal condition) and YS (yield in stress condition), we make new indices such as TOL = YP-YS or MP = (YP*YS)/2 which are highly correlated with each other. Could we perform PCA on the YP and YS variables along with indices which are making by linear or nonlinear equation from them? There are plenty of articles in Google scholar which have used these approach. In fact, we make many variables deliberately, and we reduce their dimensions using PCA!! Is it correct to do PCA for the aforementioned data?
Usually you do first PCA and then use the components in regression analyses.
Dear Akbarpour,
In my opinion principal components analysis could be applied to analyze components of Yp and Ys. Then a regression analysis could be made by using yield and yield remaining components.
- Badges
- Science topic
- Similar topics
- Biological Science
- Physiology
- Physiological Processes
- Stress