Learning with local and global consistency algorithm with a certain number of parameters, and the parameter selection of delta is sensitive about the number of iterations of the algorithm iterative process and classification results , usually by manually set of experi-ments, this approach is relatively time-consuming. In order to solve the problem,improve the classification efficiency of algorithm, this paper applies the algorithm to image classification and proposes an adaptive parameter setting method, determining the best range of the parameter delta. The experimental results show that, this paper determines the values of the parameter range can make highest classifica-tion correct rate and shortest iterative time of the algorithm;therefore this method can effectively improve the classification efficiency of algorithm.% 基于局部和全局一致性算法本身带有一定数量的参数，而参数delta的选取对算法迭代过程的迭代次数和分类结果很敏感，通常是通过实验手动设置，这种做法相对比较耗时。为了解决该问题，提高算法分类效率，文中针对该问题将算法应用到图像分类中提出了一种自适应的参数设置方法，确定参数delta的最佳取值范围。通过实验结果可以看出，确定的参数范围的取值能使算法的分类正确率最高、迭代过程所用的时间最短。因此本方法能有效地提高算法的分类效率。
展开▼
