Open krijnsent opened 4 years ago
leverage is just the trace of the hat matrix
leverage is just the trace of the hat matrix
Ehhh, you lost me a bit in the statistics lingo there... I found this explanation: https://stats.stackexchange.com/questions/208242/hat-matrix-and-leverages-in-classical-multiple-regression The explanation is quite good, but am at a loss on how to calculate it. Basically, the calculation is needed to prove that the line that was found with the regression fits okay and doesn't have too many outliers, right?
The hat matrix H= X(XTX)'XT Where X is a matrix of 2 columns length n where the first column is all 1s and the second column is lns2f_i...n T means transpose and ' means inverse. The leverage is the diagonal elements of the resulting matrix. ie leverage=Hii for i =1 to n. Leverage tells us how much influence a particular point had on the regression. We dont want to have few points with high leverage influencing our regression line.
Hi there, I took your excel file and updated the model quite a bit. Have a look: https://www.dropbox.com/s/3qlzevztkyepuja/sf2.xlsx?dl=0
Main updates:
Open questions: