We will explore the properties and methods underlying the MackChainladder class.
As usual, we we import the chainladder package as well as the popular
pandas package. For plotting purposes, we will also be using Jupyter's
%matplotlib inline magic function.
import chainladder as cl
import pandas as pd
%matplotlib inline
We will be exploring the MackChainladder class on the GenIns dataset
included in the chainladder package. Let's load the triangle and
look at it.
GI = cl.load_dataset('GenIns')
GI_tri = cl.Triangle(GI)
GI_tri.data
| dev | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
|---|---|---|---|---|---|---|---|---|---|---|
| origin | ||||||||||
| 1 | 357848 | 1124788 | 1735330 | 2218270 | 2745596 | 3319994 | 3466336 | 3606286 | 3833515 | 3901463 |
| 2 | 352118 | 1236139 | 2170033 | 3353322 | 3799067 | 4120063 | 4647867 | 4914039 | 5339085 | NaN |
| 3 | 290507 | 1292306 | 2218525 | 3235179 | 3985995 | 4132918 | 4628910 | 4909315 | NaN | NaN |
| 4 | 310608 | 1418858 | 2195047 | 3757447 | 4029929 | 4381982 | 4588268 | NaN | NaN | NaN |
| 5 | 443160 | 1136350 | 2128333 | 2897821 | 3402672 | 3873311 | NaN | NaN | NaN | NaN |
| 6 | 396132 | 1333217 | 2180715 | 2985752 | 3691712 | NaN | NaN | NaN | NaN | NaN |
| 7 | 440832 | 1288463 | 2419861 | 3483130 | NaN | NaN | NaN | NaN | NaN | NaN |
| 8 | 359480 | 1421128 | 2864498 | NaN | NaN | NaN | NaN | NaN | NaN | NaN |
| 9 | 376686 | 1363294 | NaN | NaN | NaN | NaN | NaN | NaN | NaN | NaN |
| 10 | 344014 | NaN | NaN | NaN | NaN | NaN | NaN | NaN | NaN | NaN |
To create a MackChainladder model, we can specify up to four elements. A triangle is the only non-optional element that needs to be specified to create the model. Another parameter of interest we will be using here is the alpha parameter.
The default parameter is alpha = 1
For all other parameters, please refer to the documentation of the MackChainladder class.
Load the Data
GI_mack = cl.MackChainladder(tri = GI_tri)
There are a variety of attributes and methods available in the MackChainladder class. Most of these borrow notation similar to that of the `R chainladder <https://github.com/mages/ChainLadder>`__ package, but there are a few differences. A complete list of attributes and methods are shown below. Details on these are contained in the documentation of this module.
** Available attributes and methods **
[item for item in dir(GI_mack) if item[:1]!='_']
['Fse', 'age_to_age', 'alpha', 'chainladder', 'f', 'fse', 'full_triangle', 'is_exponential_tail_appropriate', 'mack_se', 'parameter_risk', 'plot', 'process_risk', 'sigma', 'summary', 'total_mack_se', 'total_parameter_risk', 'total_process_risk', 'triangle', 'weights']
A useful method is the summary() method. This will produce, by origin period, the IBNR estimate based off of the MackChainladder model as well as its corresponding standard error. This is useful in gaining deeper insight into the uncertainty in the model.
GI_mack.summary().round(3)
| origin | Latest | Dev to Date | Ultimate | IBNR | Mack S.E. | CV(IBNR) |
|---|---|---|---|---|---|---|
| 1 | 3901463 | 1.000 | 3901463.000 | 0.000 | 0.000 | NaN |
| 2 | 5339085 | 0.983 | 5433718.815 | 94633.815 | 71835.187 | 0.759 |
| 3 | 4909315 | 0.913 | 5378826.290 | 469511.290 | 119473.736 | 0.254 |
| 4 | 4588268 | 0.866 | 5297905.821 | 709637.821 | 131572.833 | 0.185 |
| 5 | 3873311 | 0.797 | 4858199.639 | 984888.639 | 260530.015 | 0.265 |
| 6 | 3691712 | 0.722 | 5111171.458 | 1419459.458 | 410406.890 | 0.289 |
| 7 | 3483130 | 0.615 | 5660770.620 | 2177640.620 | 557795.542 | 0.256 |
| 8 | 2864498 | 0.422 | 6784799.012 | 3920301.012 | 874882.218 | 0.223 |
| 9 | 1363294 | 0.242 | 5642266.263 | 4278972.263 | 970959.785 | 0.227 |
| 10 | 344014 | 0.069 | 4969824.694 | 4625810.694 | 1362981.070 | 0.295 |
In many cases, we prefer a visual representation of the model, and can represent much of the same data contained in the summary() method by calling the plot() method.
The plot() method can be passed a list of desired plots or it can be generically called to plot all available plots.
Individual plot
GI_mack.plot(plots=['summary'])
<matplotlib.figure.Figure at 0x233ac232390>
Plotting default (all plots)
GI_mack.plot()
<matplotlib.figure.Figure at 0x233acb89cf8>
