Stochastic reserving with a stacked model based on a hybridized Artificial Neural Network

Currently, legal requirements demand that insurance companies increase their emphasis on monitoring the risks linked to underwriting and asset management activities. Regarding risks, main uncertainties insurers must manage are related premium sufficiency cover future claims adequacy of current reserves pay outstanding claims. Both calibrated using stochastic models due nature. This paper introduces a reserving model based set machine learning techniques such as Gradient Boosting, Random Forest Artificial Neural Networks. These algorithms other widely used stacked predict shape runoff. To compute deviation around former prediction, log-normal approach is combined with suggested model. The empirical results demonstrate proposed methodology can be improve performance traditional Bayesian statistics Chain Ladder, leading more accurate assessment risk.