应用多元线性回归、人工神经网络、支持向量机3种方法,对加入聚乙二醇、十二烷基苯磺酸钠、石油磺酸盐和部分水解聚丙烯酰胺四种处理剂的蒙脱土悬浮液的电动电位进行预测。在模型训练中,分别采用了神经网络集成和非启发式参数优化来提高人工神经网络和支持向量机模型的泛化能力。检验结果表明,参数优化的支持向量机模型预测精度最高,其平均误差率为3.88%,最大误差率为7.55%。 Three kinds of methods, multivariate linear regression, artificial neural network and support vector machine, were used to study the effects of four kinds of treatment agents, polyethylene glycol, sodium dodecyl benzene sulfonate, petroleum sulfonate and partially hydrolyzed polyacrylamide The electrokinetic potential of the suspension is predicted. In model training, neural network ensemble and non-heuristic parameter optimization are respectively used to improve the generalization ability of ANN and SVM models. The test results show that the parameter-optimized SVM model has the highest prediction accuracy with the average error rate of 3.88% and the maximum error rate of 7.55%.