Skip Navigation

Your Environment. Your Health.

Principal Variance Component Analysis

Workflow for PVCA Diagram Text Version

Back to PVCA main page

 

Start of diagram

 

Box Text: Preprocessed and normalized dataset (PxN)

 

Arrow flows to next box below

 

Box Text: Obtain the covariance/correlation matrix (NxN), and perform PCA analysis

 

Arrow splits and flows to two boxes, Box Text 1: Retain the eigenvalues for its corresponding principal component and Box Text 2: Determine the first few Principal Components

 

Determine the first few Principal Components flows to next box below, Box Text: Fit all effects as random factors in as a univariate mixed model to each principal component, variance component estimates were retained

 

Box Retain the eigenvalues for its corresponding principal component rejoins the flow here

 

Arrow flows to next box below

 

Box Text: For each estimate, compute the weighted average from each REML estimate based on the eigenvalues retained from the principal component analysis

 

Arrow flows to next box below

 

Standardize averaged batch effects across all estimates include the model residual, to proportion of total variance

 

End of diagram

 

Back to PVCA main page

Back to Top