getAUC: Calculates AUC for a logistic regression.

In buddha2490/BERGMets: Metabolomics data and analysis in the BERG

Description

getAUC() takes a series of outcomes, metabolites, and covariates and calculates AUC for each model. You can input as many outcomes and metabolites as you want.

Usage

getAUC(dat, outcome, compid, covariates=NULL, normalize=T)

Arguments

dat	Data frame containing your metabolites, outcome, and covariates
outcome	Numeric vector coded [0,1] for your outcome.
compid	Vector of COMP_IDs
covariates	Vector of covariate variable names (optional)
normalize	Normalize the data prior to modeling? (default is TRUE)

Value

If a single outcome is provided, getAUC() will return a data frame with the following variables

COMP_ID	Each COMP_ID listed in compid argument
OUTCOME	Outcome
AUC

AUC calculation

If multiple outcomes are requested, getAUC() will return a list. Each data frame of the list corresponds to each of the outcomes.

Author(s)

Brian Carter and Becky Hodge

See Also

roc, normalizeMets,

Examples

# load metabolite data

metabolites <- getMetabolites("breast_metabolomics")
df <- left_join(survey,metabolites$metabolites,"ID")

comp.id <- metabolites$biochem$COMP_ID[1:10]

# code BMI categories as [0,1] - obese vs normal
df$bmi_binary <- ifelse(df$BMI < 25,0, ifelse(df$BMI >= 30,1,NA))

myAUC <- getAUC(df,
                outcome="bmi_binary",
                compid = comp.id,
                covariates="AGE_INT",
                normalize=T)

buddha2490/BERGMets documentation built on Sept. 6, 2020, 5:11 p.m.