| 基于EM的朴素贝叶斯分类算法 |
| |
| 引用本文： | 张亚萍,陈得宝,侯俊钦.基于EM的朴素贝叶斯分类算法[J].宿州学院学报,2010,25(11). |
| |
| 作者姓名： | 张亚萍 陈得宝 侯俊钦 |
| |
| 作者单位： | 淮北师范大学物理与电子信息学院 |
| |
| 基金项目： | 安徽省高等学校省级优秀青年人才基金，安徽省自然科学基金 |
| |
| 摘 要： | 将EM算法引入到朴素贝叶斯分类研究中,提出一种基于EM的朴素贝叶斯分类算法。首先用未缺失的数据属性的算术均数作为初始值,求得极大似然估计;其次迭代执行算法的E步和M步直至收敛,然后完成缺失数据的填补;最后根据朴素贝叶斯分类算法对数据进行分类。实验结果表明,与朴素贝叶斯分类算法相比,基于EM的朴素贝叶斯分类算法具有较高的分类准确率。
|
| 关 键 词： | 朴素贝叶斯分类 先验概率 后验概率 EM算法 缺失数据 |
Naive Bayesian Classification Based on EM Algorithm |
| |
| ZHANG Ya-ping,CHEN De-bao,HOU Jun-qin.Naive Bayesian Classification Based on EM Algorithm[J].Journal of Shuzhou College,2010,25(11). |
| |
| Authors: | ZHANG Ya-ping CHEN De-bao HOU Jun-qin |
| |
| Institution: | ZHANG Ya-ping; CHEN De-bao; HOU Jun-qin(Physics and Electronic Information Institute; Huaibei Teacher′s University; Huaibei; Anhui; 235000; China); |
| |
| Abstract: | A Naive Bayesian classification based on EM algorithm is researched.EM algorithm was used to calculate the principal based on incomplete data,maximum likelihood estimation.Each iterative algorithm includes two steps: the first step in seeking expectations(Expectation Step),known as the E step;the second step for maxima(Maximization Step),known as step-by-step M,then the absent value of the record is filled by the corresponding attribute of the data.Finally,the handled data set is clustered by naive Bayesian... |
| |
| Keywords: | nave bayesian classification prior probability posterior probability EM algorithm missing data |
| 本文献已被 维普 万方数据 等数据库收录！ |
