中文摘要：

为提高变形抗力的预测精度,提出了一种基于混合最小二乘支持向量机和数学模型的组合方法。在该方法中,最小二乘支持向量机的参数通过基于退火策略的自适应粒子群优化算法自动获得。仿真实验结果表明,该组合方法不仅能够重现样本数据的变形抗力,还能非常精确地预测非样本数据。通过与其它文献中常用方法的比较发现,该方法在变形抗力预测的有效性和精确性方面都有很大提高。

英文摘要：

To improve the prediction accuracy of the flow stress, a hybrid model based on the Hybrid Least Squares Support Vector Machine （HLS-SVM） and Mathematical Models （MM） was proposed. In HLS-SVM model, the optimal parameters of LS-SVM were obtained by self-adaptive Particle Swarm Optimization （PSO） based on Simulated Annealing （SA）. Simulation experiment results revealed that this model could correctly recur to the flow stress in the sample data and accurately predict the non-sample data. The efficiency and accuracy of the predicted flow stress achieved by the proposed model were better than the methods used in most literature.