Skip to main content

kernels.fit_conventional()

Fit one cohort using only information stored on or before as_of.

Usage

Source

kernels.fit_conventional(
    triangle,
    candidate,
    *,
    as_of,
    loss_field="paid_loss",
    premium_field="earned_premium"
)

Premium is read from the same historical slice. BF’s a priori loss ratio is supplied by the caller: it must have been available then too. Negative increments are permitted; cumulative amounts must be finite/non-negative.