Capping and Trending

[henrydingliu]

The tutorials go through on-the-fly capping and trending directly on the Triangle class. Is there a way to apply detrended caps?

[jbogaardt]

Maybe like this?

import chainladder as cl
import numpy as np

prism = cl.load_sample('prism')['Paid'].incr_to_cum()
detrended_cap = (prism.sum().latest_diagonal*0+25000).trend(-.05) # Create a detrended cap
detrended_cap_triangle = np.minimum(prism, detrended_cap) # Apply the cap