一种不确定数据的朴素贝叶斯分类方法

摘要

Aiming at the classification problem with uncertain data, a naive Bayes classification method is proposed in this paper-uncertainty naive Bayes classifier based on histogram estimation (HU-NBC). Based on the idea of histogram estimation, an estimation model of novel probability density functions is established for uncertain data and is used to estimate the class-conditional probability density functions of uncertainty naive Bayes classifier. Experimental results on UCI datasets show that HU-NBC has a good classifying accuracy, less runtime and memory requirements compared with existing methods, and it is suitable for classification with large amounts of data.%针对不确定性数据的分类问题，提出一种基于直方图估计的不确定性朴素贝叶斯分类器（HU-NBC）。基于直方图估计的思想，建立估计不确定性数据概率密度函数的数学模型，并利用该模型估计不确定性朴素贝叶斯分类器的类条件概率密度函数。实验结果表明，与同类型算法相比，基于直方图估计的HU-NBC算法拥有较优的分类精度、较小的时间代价和空间需求，适合解决数据量较大的不确定性数据分类问题。