Clark Growth Curves

import chainladder as cl
import numpy as np

This example demonstrates one of the attributes of the ClarkLDF. We can use the growth curve G_ to estimate the percent of ultimate at any given age.

# Grab Industry triangles
clrd = cl.load_sample("clrd").groupby("LOB").sum()

# Fit Clark Cape Cod method
model = cl.ClarkLDF(growth="loglogistic").fit(
    clrd["CumPaidLoss"], sample_weight=clrd["EarnedPremDIR"].latest_diagonal
)

# sample ages
ages = np.linspace(1, 300, 30)

# Plot results
results = model.G_(ages).T

Show code cell source

Hide code cell source

import matplotlib.pyplot as plt

plt.style.use("ggplot")
%config InlineBackend.figure_format = 'retina'

ax = results.plot(
    title="Loglogistic Growth Curves", xlabel="Age", ylabel="% of Ultimate"
);