Towards Predicting Neural Net Control of Macro Econometric Multi Compartment Models

In this paper we propose a novel concept for analyzing dynamic economic and nan cial systems by multi compartment mod eling techniques in combination with neu ral networks This is done as a two step process rst a multi layer percep tron MLPident is introduced to approxi mate the dynamics of observable variables of a single domestic compartment and to identify interdependencies second a multi layer perceptron MLPcontrol evaluates se ries of outputs from MLPident to make a complex decision about certain properties e g stability and to nd a classi cation of economic or nancial scenarios Simula tion results con rm the ability of MLPident to learn the structure and parameters of the domestic compartment and to provide good results in out of sample tests Simulation results for MLPcontrol show that it is able to decide whether a current parameter set leads to a stable or instable constellation in the future Possible applications of this novel concept are analysis of interdependen cies of economic variables between multiple countries or control of dynamic system be haviour with regard to predicting proper ties