Does anyone know of a function that turns the coefficients of lm( y~ poly(x1 , x2, degree=2)) into something that can be interpreted easily. I was think along the lines of the matrix representation of quadratic forms: (x-mu)'A(x-mu) +k , and finding the eigenvectors/values of A, and the vector mu, but anything that allows me to visualise a contour plot would be great. Thanks Simon Bond
Possibly Parallel Threads
- interpret significance from the contr.poly() function
- poly() can exceed degree k - 1 for k distinct points (PR#11251)
- Issue with prediction from lm object with poly
- poly() in lm() leads to wrong coefficients (but correct residuals)
- example to demonstrate benefits of poly in regression?