MergeGUI (version 0.2-1)

scale_rpart: Compute the misclassification rate for each variable. When merging data from several datasets, it is meaningful to detect whether the matched variables from different files have different centers. The function computes the misclassification rate variable by variable using classification tree (the rpart package). It will firstly merge the dataset by the given nametable.class, then use rpart for each variable to seperate the data without any covariates and compute the misclassification rate.

Description

Compute the misclassification rate for each variable. When merging data from several datasets, it is meaningful to detect whether the matched variables from different files have different centers. The function computes the misclassification rate variable by variable using classification tree (the rpart package). It will firstly merge the dataset by the given nametable.class, then use rpart for each variable to seperate the data without any covariates and compute the misclassification rate.

Usage

scale_rpart(nametable.class, dataset.class, name.class, varclass = NULL)

Arguments

nametable.class

A matrix of the matched variable names. The number of columns is equal to the number of files. The column names are required. Each row represents a variable that is going to be merged. Any elements except NA in nametable.class must be the variable names in dataset.class.

dataset.class

The dataset list. The length of the list is equal to the number of files, and the order of the list is the same as the order of columns in nametable.class.

name.class

A character vector of variable names. The length of the vector must be equal to the number of rows in nametable.class. Since the variable names in nametable.class may not be consistent, name.class is needed to name the variables.

varclass

A character vector of variable classes. The length of the vector must be equal to the number of rows in nametable.class. All the classes should be in "numeric", "integer", "factor", and "character". Default to be null, then it will be determined by var.class.

Value

A vector of the misclassification rate. The rate is between 0 and 1, or equal to 9 if one of more groups only have NA's.

Examples

a = data.frame(aa = 1:5, ab = LETTERS[6:2], ac = as.logical(c(0, 1, 0, NA, 0)))
b = data.frame(b1 = letters[12:14], b2 = 3:1)
dat = list(a, b)
name = matrix(c("ab", "aa", "ac", "b1", "b2", NA), ncol = 2)
colnames(name) = c("a", "b")
newname = c("letter", "int", "logic")
scale_rpart(name, dat, newname)