Split modeling data into test and train set

Share:

Description

Takes in data, fraction (for train set) and seed, and returns train and test set

Usage

1
splitdata(data, fraction, seed = NULL)

Arguments

data

a matrix, data.frame or data.table

fraction

proportion of observations that should go in the train set

seed

an integer value

Details

An essential task before doing modeling is to split the modeling data into train and test sets. splitdata is built for this task and returns a list with train and test sets, which can be picked using the code given in example.

fraction corresponds to the train dataset, while the rest of the observations go to the test dataset. If the user wants to generate the same test and train dataset everytime, he should specify a seed value.

Value

a list with two elements: train and test set

Author(s)

Akash Jain

See Also

actvspred, mape, accuracy, auc, iv, ks

Examples

1
2
3
4
5
6
7
8
9
# A 'data.frame'
df <- data.frame(x = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10),
                 y = c('a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j'),
                 z = c(1, 1, 0, 0, 1, 0, 0, 1, 1, 0))

# Split data into train (70%) and test (30%)
ltData <- splitdata(data = df, fraction = 0.7, seed = 123)
trainData <- ltData$train
testData <- ltData$test