论文部分内容阅读
支持向量机(Support Vector Machine,SVM)是以核函数为载体的机器学习方法,集优化、核、最佳推广能力等特点于一身,目前比较广泛的一个应用是数字图像分类,具体的步骤是:先用词袋模型对数字图像特征进行组织,再构造核函数进行训练、学习,然后分类。在整个过程中,对最后分类结果起到关键作用的分别是核函数的构造和分类器核参数的选择,为解决核参数大多依靠经验选取或者大范围网络搜索耗时等问题,引入群智能算法来优化核参数,使得模型性能达到最优。最后选用Caltech 101、Caltech 256中的经典图像数据集做分类实验,以验证其核参数优化方法的有效性。
Support Vector Machine (SVM) is a machine learning method based on kernel function. It combines optimization, kernel and best promotion ability. Currently, one of the most widely used applications is digital image classification. The specific steps are: : The first use of the bag model of digital image features to organize, and then construct the kernel function training, learning, and then classified. In the whole process, the key functions of the final classification result are the construction of the kernel function and the choice of the kernel parameters of the classifier. In order to solve the problems that most of the kernel parameters rely on experience selection or large-scale network search, the group intelligence algorithm To optimize the nuclear parameters, making the model to achieve the best performance. At last, we use the classic image dataset from Caltech 101 and Caltech 256 to do classification experiments to verify the validity of the method of kernel parameter optimization.