Regression Flashcards
Formula for converting from probability to logit
logit(y) = log(p(y)/1-p(y))
What is the inverse function to transform from logit to probability
p = exp(y)/1+exp(y)
what is the link function for a gaussian (normal) distribution
identity
what is the link function for a binomial distribution
logit
how to centre a predictor
column - mean(column)
what is the intercept when predictors are centred in multiple regression
the grand mean of the DV
how to compare multiple regression models
using ANOVA or comparing adjustedR2
what is name of the curve produced by plotting the slope of B1 in a binomial regression
signoid curve
what does applying a linear function to the logit of p(y) produce
a simple linear slope and intercept
what is the model comparison criteria for binomial logistic regression using glm()
akaike information criteria
what number do we want the akaike information criteria to be closer to
0
what does the incercept represent in a binomial logistic regression with centred predictors
the log odds of the probability of falling into one category
what kind of regression would be use a cumulative logit model function for
Ordinal scale DVs
What is the depedent variable in regression
the outcome variable
how many intercept estimates is that for a rating acale with 6 points in an ordinal regression
5
