A Neural Network Linking Process for Insurance Claims

A novel method of integrating multiple neural networks into one large network via a linking process is used to combine neural networks for the purpose of Insurance Claim Reservation. Neural networks are commonly trained to solve a specific problem for an encapsulated problem domain. Simple problems can be solved by a single network, whereby more complicated problems can be solved by sub-networks. These sub-networks then re-combined via a linking process forming a combined network, which has the ability to solve the entire problem. Insurance companies require monetary reserves for accounting, calculation of premium, reinsurance and asset liability management, which includes claims payable. Since the business of insurance companies lies in the future, accurate claims estimation for future financial years will increase profitability and ensure their solvency. A fine balance between under-reserving and over-reserving must be found because under-reserving can threaten solvency and over-reserving reduces profitability. Therefore this novel method for claims estimation with neural network linking has been developed.