论文部分内容阅读
针对远海任务舰船备件器材的分类管理,通过运用支持向量机理论,充分发挥多尺度核在非线性分类中的优势,借助最小二乘原理,构建了多尺度最小二乘支持向量机学习模型。在实际运用中,通过选用高斯径向函数作为多尺度核函数,以训练样本数据分布的离散系数作为核函数宽度参数取值依据,采取ECOC方法建立了多类分类模型,实例计算表明,该方法对远海任务舰船备件器材的分类是有效、可行的。
Aiming at the classification and management of spare parts and equipment for distant missions, the multi-scale least square support vector machine learning model is constructed by using the principle of least squares with the help of support vector machine theory, giving full play to the advantages of multi-scale kernel in non-linear classification. In practical application, by using the Gaussian radial function as the multi-scale kernel function and the discrete coefficient of the training sample data distribution as the basis of the kernel function width parameter, the ECOC method is used to establish a multi-class classification model. The calculation results show that this method The classification of offshore spare parts for ships and ships is effective and feasible.