Stochastic loss reserving with mixture density neural networks

@article{AlMudafer2021StochasticLR,
  title={Stochastic loss reserving with mixture density neural networks},
  author={Muhammed Taher Al-Mudafer and Benjamin Avanzi and Greg Taylor and Bernard Wong},
  journal={Insurance: Mathematics and Economics},
  year={2021}
}
2 Citations

Ensemble Distributional Forecasting for Insurance Loss Reserving

This paper proposes a systematic framework to objectively combine (i.e. ensemble) multiple stochastic loss reserving models such that the strengths ofiered by different models can be utilised e⬀ectively.

SPLICE: a synthetic paid loss and incurred cost experience simulator

In this paper, we first introduce a simulator of cases estimates of incurred losses called SPLICE (Synthetic Paid Loss and Incurred Cost Experience). In three modules, case estimates are simulated

References

SHOWING 1-10 OF 48 REFERENCES

The Actuary and IBNR Techniques: A Machine Learning Approach

It is shown that the use of optimal techniques can lead to more accurate reserves and investigate the circumstances under which different scoring metrics should be used.

Artificial Neural Networks in Insurance Loss Reserving

This paper uses Artificial Neural Networks (ANN) for the problem of Loss Reserving, the practice of estimating the future payments for the claims which have occurred on an insurance portfolio.

Neural network embedding of the over-dispersed Poisson reserving model

To embed a classical actuarial regression model into a neural network architecture, the cross-classified over-dispersed Poisson model for general insurance claims reserving is considered and it is demonstrated how this model can be improved by neural network features.

When Networks Disagree: Ensemble Methods for Hybrid Neural Networks

Experimental results show that the ensemble method dramatically improves neural network performance on difficult real-world optical character recognition tasks.

AI in actuarial science – a review of recent advances – part 1

How actuarial science may adapt and evolve in the coming years to incorporate these new techniques and methodologies based on a modern approach to designing, fitting and applying neural networks, generally referred to as “Deep Learning.”

Using Machine Learning to Model Claims Experience and Reporting Delays for Pricing and Reserving

It is shown how GLMs, lasso regression, gradient boosted trees and deep learning can be applied to the new formulation of mortality incidence models to produce results of superior accuracy compared to the traditional approaches.

Penalising Unexplainability in Neural Networks for Predicting Payments per Claim Incurred

A granular machine learning model framework to jointly predict loss development and segment risk pricing is presented and generally outperformed PPCI on predicting ultimate loss for sufficient sample size.

DeepTriangle: A Deep Learning Approach to Loss Reserving

A novel approach for loss reserving based on deep neural networks is proposed that allows for joint modeling of paid losses and claims outstanding, and incorporation of heterogeneous inputs, and can be automated to produce forecasts more frequently than manual workflows.

Neural networks applied to chain–ladder reserving

This work starts from a coarse reserving method, which in this case is Mack’s chain–ladder method, and shows how this approach can be refined for heterogeneity and individual claims feature information using neural networks.