Improvement on Generalization Ability of the ANN-Bidding System with Genetic Algorithm

摘要

The paper deals with the problem of bidding estimation. On the foundation of successful bidding experiences in the past, a bidding system is built up based on the artificial neural network (ANN) technology and is optimized using Genetic Algorithm (GA). This not only solves the difficulty of identifying the form of function correctly but also improves the generalization ability of the system. The generalization error of the testing data is decreased to 4.0%, which is much smaller than traditional artificial neural networks' (ANN).