StepReg (version 1.4.2)

stepwiselogit: Stepwise Logistic Regression

Description

Stepwise logistic regression analysis selects model based on information criteria and Wald or Score test with 'forward', 'backward', 'bidirection' and 'score' model selection method.

Usage

stepwiselogit(data, y, exclude = NULL, include = NULL, selection = "bidirection",
  select = "SL", sle = 0.15, sls = 0.15)

Arguments

data

Data set including dependent and independent variables to be analyzed

y

A character or numeric vector indicating the subset of dependent variables

exclude

A character or numeric vector indicating the subset of independent variables removed from logstic stepwise regression analysis

include

Forces the effects vector listed in the data to be included in all models. The selection methods are performed on the other effects in the data set

selection

Model selection method including "forward", "backward", "bidirection" and 'score',forward selection starts with no effects in the model and adds effects, backward selection starts with all effects in the model and removes effects, while bidirection regression is similar to the forward method except that effects already in the model do not necessarily stay there, and score method uses information criteria to find a specified number of best models containing one, two, or three variables, and so on, up to the single model containing all of the explanatory variables.

select

Specifies the criterion that uses to determine the order in which effects enter andr leave at each step of the specified selection method including Akaike Information Criterion(AIC), the Corrected form of Akaike Information Criterion(AICc),Schwarz criterion(SBC), and Significant Levels(SL)

sle

Specifies the significance level for entry, default is 0.15

sls

Specifies the significance level for staying in the model, default is 0.15

Value

RegressionModelsSelectedbyInformationCriterion

summary of regression models selected by information criterion

SummaryOfSelection

summary of selection process

AnalysisOfMaximumLikelihoodEstimate

analysis of maximum likelihood estimate for the selected model

GoodnessOfTeste

Hosmer and Lemeshow goodness of fit (GOF) test

References

Agresti, A. (1984), Analysis of Ordinal Categorical Data, New York: John Wiley & Sons.

Agresti, A. (2014), Categorical Data Analysis, New York: John Wiley & Sons.

Hurvich, C. M. and Tsai, C.-L. (1993), A Corrected Akaike Information Criterion for Vector Autoregressive Model Selection, Journal of Time Series Analysis, 14, 271-279

Hosmer, D. W., Jr. and Lemeshow, S. (2000), Applied Logistic Regression, 2nd Edition, New York: John Wiley & Sons.

Examples

# NOT RUN {
  set.seed(1)
  yd <- data.frame(sample(c(0,1),30,replace=TRUE))
  colnames(yd) <- "remiss"
  set.seed(4)
  xd <- data.frame(matrix(c(round(rnorm(100,0,2),2),round(rnorm(140,2,4),2),
  sample(c(1,0),30,replace=TRUE),sample(1:80,30,replace=TRUE)),30,10))
  colnames(xd) <- c(paste("X",1:8,sep=""),"gender","age")
  yx <- cbind(yd,xd)
  y <- "remiss"
  stepwiselogit(yx,y,selection="bidirection",select="IC(3/2)")
# }