train_val_split_method2 {ADLP}

Train-Validation Split by Accident Period Method 2

Description

Function for training/validation splitting.

Usage

train_val_split_method2(df, tri.size, val_ratio, test = FALSE)

Arguments

df

Claims Triangle and other information. data.frame format of claims and related information for each cell. Dataframe will have columns origin and dev as columns 1 and 2 respectively.

tri.size

Triangle size.

val_ratio

Value between 0 and 1 as the approximate size of validaiton set.

test

Returns the test set if TRUE .

Details

Approximates the validation set by defining the training set as the cells below the function ((b^{1/a} - x^{1/a})^a). Where b is equal to the triangle size and a is optimised to best fit val_ratio.

The training set is therefore cells outside of this period but within the upper triangle. The test set is all observations in the lower triangle.

Note that accident period 1 and development period 1 will always be within the training set.

Value

List containing ⁠$train⁠, ⁠$valid⁠, ⁠$test⁠, which should partition the input df.

See Also

train_val_split

Examples

data("test_claims_dataset")

train_val <- train_val_split_method1(
    df = test_claims_dataset,
    tri.size = 40,
    val_ratio = 0.3,
    test = TRUE
)