computeR2: Compute variance explained by models

In HTRX: Haplotype Trend Regression with eXtra Flexibility (HTRX)

Compute variance explained by models

Description

Compute the variance explained by a linear or generalized linear model.

Usage

mypredict(model, newdata)

computeR2(pred, outcome, usebinary = 1)

Arguments

model	a fitted model, which is the output of themodel.
newdata	a data frame which contains all the variables included in the model. This data frame is used to make prediction on.
pred	a vector of the predicted outcome.
outcome	a vector of the actual outcome.
usebinary	a non-negative number representing different models. Use linear model if usebinary=0, use logistic regression model via fastglm if usebinary=1 (by default), and use logistic regression model via glm if usebinary>1.

Details

The variance explained by a linear model is based on the conventional R². As for logistic regression, we use McFadden's R².

Value

mypredict returns a vector of the predicted outcome.

computeR2 returns a positive number of the variance explained by the linear model (conventional R²) or the generalized linear model (McFadden's R²).

References

McFadden, Daniel. "Conditional logit analysis of qualitative choice behavior." (1973).

Examples

## create datasets
x=matrix(runif(100,-2,2),ncol=5)
outcome=(0.5*x[,2] - 0.8*x[,4] + 0.3*x[,5])>runif(100,-2,2)

## create binary outcome
outcome[outcome]=1
data=data.frame(outcome,x)

## compute the variance explained by features
model=themodel(outcome~.,data[1:80,],usebinary=1)
outcome_predict=mypredict(model,data[81:100,])
computeR2(outcome_predict,data[81:100,'outcome'],usebinary=1)

HTRX documentation built on May 29, 2024, 5:53 a.m.