采用快速K近邻与Kmeans聚类算法来计算前K个近邻，舍弃了一部分不可能成为待测样本的前K个近邻的训练样本，从而减少了计算量，提高了分类速度-Fast K-nearest neighbor Kmeans clustering algorithm to calculate the K nearest neighbors, abandoning the training samples of the part can not become the first K neighbors of the t