Generation of Ultimate & IBNR based on BF Method #246
Replies: 3 comments 2 replies
-
Were you having trouble following the tutorial here? |
Beta Was this translation helpful? Give feedback.
-
It is possible. One just needs to exploit the fact that the underlying formulation of the BF apriori ultimate is comauto = cl.load_sample("clrd").groupby("LOB").sum().loc["comauto"]
# Ignore apriori rate
bf_model = cl.BornhuetterFerguson(apriori=1.0)
# Instead do everything in the sample_weight
sample_weight = (comauto["EarnedPremNet"].latest_diagonal*.75).trend(.05)
bf_model.fit(comauto["CumPaidLoss"], sample_weight=sample_weight) |
Beta Was this translation helpful? Give feedback.
-
I don't understand the question. Trend is an example, but not necessary for all use-cases. The point that should be understood is that neither the |
Beta Was this translation helpful? Give feedback.
-
I hope that you are fine and perfect.
I would like to ask you that what should I do for generation of Ultimate & IBNR based on the BornhuetterFerguson (BF) and how to put the bf rate.
Please help me in this regards.
Beta Was this translation helpful? Give feedback.
All reactions