Outputting risk groups for a logistic regression model


I have a problem with outputting the terms for a logistic regression model in R. For a given list of independent values, say list l of terms {w,y,z} to determine dependent variable {x}, I want to find out what the biggest regressor is when we pair two terms together. I want to be able to group multiple independent variables together and say "when a record has this combination of values, then they have a very strong chance of predicting X". I tried to just add the interactions when calling the glm function like glm(x~y + w + z + w:z + y:z + y:w, data = l). But the results come out very hard to explain, because they of how they are measured between themselves and not just measured against the mean. Does anyone know a way to do this?


Posted 2016-03-09T17:10:16.333

Reputation: 133

No answers