Transmission Theory of the Risk Neural Network
Cunbin Li, Kecheng Wang · 2007 IFIP International Conference on Network and Parallel Computing Workshops (NPC 2007) · 2007
On the basis of general project risk element transmission theory, in order to consider an effective solution of portfolio selection problem. This research proposed a new neural network model: risk neural network model. Based on artificial neural network model analysis, we introduced three computational rules: analysis, weighted and polymerization in risk neural network. By solving the hidden layer of random variables characteristic function, a series of definitions were built in risk neural network, including feed-forward risk neural network model and muti-hidden layer risk neural network model. According to the series of definitions, we divided risk elements into discrete model and continuous model to be discussed separately, eventually the analytical model of risk neural network was built. The interactive approach was illustrated with a practical example.