The Application of Macro-Economic Prediction Based on Improved Gene Expression Programming

An improved gene expression programming (IGEP) is proposed in this paper. It has some new features: 1) introducing a new individual coding; 2) introducing a new way of creating constants; 3) introducing a hybrid self-adaptive crossover-mutation operator, which can enhance the search ability and exploit the optimum offspring. To validate the performance of IGEP, this paper applies IGEP into the solution of the macro-economic predictions. The experimental results demonstrate that Improved GEP can automatically find better Optimization Model, based on which prediction will be generated much more exactly.