Logistic Regression Notes

j proportional odds model log i 2 j j 1

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: nes an ordinal scale on j . UNM Cummulative Logit model Cummulative odds for the j − th category. πi + π2 + . . . + πj P (Z ≤ cj ) = P (Z > cj ) πj +1 + πj +2 + . . . + πJ Proportional odds model: log πi + π2 + . . . + πj πj +1 + πj +2 + . . . + πJ = β0j + β1 X1 + . . . + βp Xp only intercept β0j depends on j . β1 , β2 , . . . , βp are constant across j . UNM Alternatively, consider ratios πj −1 π1 π2 , ,..., ,... π2 π3 πj Adjacent category logic model log πj πj +1 = β0j + β1 X1 + . . . + βp Xp or log πj πj +1 + πj +2 . . . + πK = β0j + β1 X1 + . . . + βp Xp Odds of being in category j (cj −1 ≤ Z ≤ cj ), conditional on Z > cj −1 . Other link functions can be considered. UNM...
View Full Document

This note was uploaded on 01/27/2014 for the course STAT 574 taught by Professor Gabrielhuerta during the Fall '13 term at New Mexico.

Ask a homework question - tutors are online