为进一步改善局部支持向量机的分类效率和分类精度,提出一种改进的局部支持向量机算法。该算法对每类训练样本分别进行聚类,使用聚类生成的样本中心点集代替样本,使用改进的k最近邻算法选取测试样本的k个近邻。分别在UCI数据集和自建树皮图像数据集上对本研究算法的有效性进行测试。实验结果表明,本研究提出的算法在分类精度和效率上具有一定的优势。 In order to further improve the classification efficiency and classification accuracy of local support vector machines, an improved local support vector machine algorithm is proposed. The algorithm clusters each type of training samples separately, uses the set of sample center points generated by clustering instead of samples, and selects k nearest neighbors of the test samples by using improved k nearest neighbors algorithm. The validity of the proposed algorithm is tested on UCI datasets and self-built bark image datasets respectively. Experimental results show that the proposed algorithm has certain advantages in classification accuracy and efficiency.