说明：  C针对模式识别问题H描述了支持向量机的基本思想H着重讨论了OD=?PI最小二乘=?PI加权=?P 和直接 =?P 等新的支持向量机方法H用于降低训练时间和减少计算复杂性的海量样本数据训练算法分块法I分解法H提 高泛化能力的模型选择方法H以及逐一鉴别法I一一区分法IPD., 分类法I一次性求解等多类别分类方法@最后给 出了污水生化处理过程运行状态监控的多类别分类实例@作为结构风险最小化准则的具体实现H支持向量机具有 全局最优性和较好的泛化能力
(C for pattern recognition problem of H describes the support vector machine basic idea of H focused on the OD =? PI squares =? PI weighted =? P and direct =? P, such as new support vector machine method used to reduce training time H and reduce the computational complexity of the mass sample data training algorithm block decomposition H Act I improve generalization ability of the model selection method to identify H as well as one by one law I 11 to distinguish between law IPD., classification, etc. I one-time solution of multi-class classification method @ Finally, biological wastewater treatment processes to monitor the operational status of the multi-class classification example @ as structural risk minimization criteria for the concrete realization of H support vector machine with the global optimum and good generalization ability)