MackNet_Boot_Incurred: MackNet_Boot_Incurred

In EduardoRamosP/MackNet: Mack-Net model: Blending Mack's model with Recurrent Neural Networks

Description

It produces the boostrapping to obtain a full reserves distribution with the MackNet model.

Usage

MackNet_Boot_Incurred(
  Simulations,
  Parameters,
  Cumulative.T,
  Cumulative.T.Incurred,
  OutliersControl
)

Arguments

Simulations	Number of samples produced by the MackNet model
Parameters	List with all the MackNet parameters. Please refer to function MackNet_Parameters for further details
Cumulative.T	Cumulative payments triangle.
Cumulative.T.Incurred	Incurred cost triangle.
OutliersControl	Development factors below 0.975 are not allowed when this variable is set to 1.

Value

The formula generates the following outputs:

• Residuals_Boot Residuals used within the bootstrap procedure.

• Payments.D It contains the total following year payments sampled by the MackNet model.

• Reservest.D It contains the total reserves sampled by the MackNet model.

• Reservest_1.D It contains the total reserves in t+1 sampled by the MackNet model.

• Ultimatest.D It contains the ultimate cost sampled by the MackNet model.

• Payments.D.Vector It contains the following year payments by accident year sampled by the MackNet model.

• Reservest.D.Vector It contains the reserves by accident year sampled by the MackNet model.

• Reservest_1.D.Vector It contains the reserves in t+1 by accident year sampled by the MackNet model.

• Ultimatest.D.Vector It contains the ultimate cost by accident year sampled by the MackNet model.

• TrianglesBackup This array contains all the triangles sampled by the MackNet model.

• UpperTrianglesBackup Only for testing purposes. It assumes that the only observed value is the one included in the first development year.

• Boot_DevFactors Predictive development factors sampled by the boostrap procedure.

• Development_Factors_MackNet Predictive development factors of the MackNet model.

• Alpha Predictive alpha by development year of the MackNet model. This variable is named as sigma in "Mack-Net model: Blending Mack's model with Recurrent Neural Networks".

EduardoRamosP/MackNet documentation built on Sept. 26, 2020, 9:21 a.m.