Abstract: The support vector regression(SVR)approach based on the quantum-behaved particle swarm optimization(QPSO)for its parameters optimization was proposed to predict the contribution degrees to the metallogenesis.The results show that the prediction precision of SVR method is superior to that of BP neural networks.It is suggested that SVR is an effective and powerful tool for predicting the contribution degrees to the metallogenesis.

Key words: Metallogenesis prediction, Support vector machines(SVM), Quantum-behaved particle swarm optimization, BP neural network