：将K—means算法引入到朴素贝叶斯分类研究中，提出一种基于K—means的朴素贝叶斯分类算法。首先用K—

me．arks算法对原始数据集中的完整数据子集进行聚类，计算缺失数据子集中的每条记录与 个簇重心之间的相似度，把记

录赋给距离最近的一个簇，并用该簇相应的属性均值来填充记录的缺失值，然后用朴素贝叶斯分类算法对处理后的数据

集进行分类。实验结果表明，与朴素贝叶斯相比，基于K—means思想的朴素贝叶斯算法具有较高的分类准确率。-: K-means algorithm will be introduced to the Naive Bayesian Classifier study, a K-means based on the Naive Bayesian classification algorithm. First of all, with K-me. arks algorithm focus on the raw data of the complete data subset of the cluster, the calculation of missing data for each subset of records and the similarity between the cluster center of gravity to the nearest record assigned to a cluster, and the corresponding attributes of the cluster means to fill the missing value record, and then use Naive Bayes classification algorithm to deal with the data set after classification. The experimental results show that compared with the Naive Bayes, K-means based on the thinking of Naive Bayes algorithm has higher classification accuracy.