univ: univariate calcualted for data framework

In billyuanyao/WOECredit:

Description

univariate calculation for all independent variables in a data framework mean, median, variance, std, missing rate, unique rate

Usage

univ(Data, keeplist = NULL, intmiss = NULL)

Arguments

Data	data frame with at least two columns
keeplist	Name of the Independent Variables keept for capping, if missing then for all Independent Variables
intmiss	automatically fill missing values. defuat for numarical vairable is NA, but can be 0 (there would be multiple missing types of a model)

Value

test: data is from the Titanic project https://www.kaggle.com/c/titanic/data traindata <- read.csv('train.csv',header=T,na.strings=c("")) Data <- subset(traindata,select=c(2,3,5,6,7,8,10,12)) univ(Data)

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (Data, keeplist = NULL, intmiss = NULL)
{
    if (!is.null(keeplist)) {
        Data <- Data[, keeplist]
    }
    if (is.null(intmiss)) {
        intmiss <- 0
    }
    nums <- sapply(Data, is.numeric)
    Data <- Data[, nums]
    varlist <- sapply(Data, function(x) {
        data.frame(mean = mean(x), median = median(x), var = var(x),
            sd = sd(x), nmiss = sum(is.na(x)), n = length(x),
            missrate = sum(is.na(x))/length(x))
    })
    transfvar <- data.frame(t(varlist))
    nums <- sapply(Data, is.numeric)
    filldata <- Data[, nums]
    filldata[is.na(filldata)] <- intmiss
    Data <- filldata
    return(transfvar)
  }

billyuanyao/WOECredit documentation built on May 28, 2019, 7:11 p.m.