I analysed a data set of 375 observations (genotypes) with 52 quantitative and qualitative variables through PCA based on correlation matrix. I got 17 components with eigen value more than 1. The first PC was having only 9 per cent variation. I felt it was very low. I didn't do any transformation od data before PCA.
Is there any need to do data transfermation even if we use correlation matrix. Suggest what type of data transformation can be done to reduce the dimensions.
Sir, Before PCA analysis, log transformation can be used for make non-linear to linear data set. Generally, sum of Eigenvalues is equal to number of characters from all PCs. In your case 17 components comes from all PCs. So Log transformation can be helpful. and check for Multicollinearity.
- Badges
- Science topic
- Similar topics
- Mathematics
- Algebra
- Matrix