处理缺失数据的朴素贝叶斯分类增量算法

Incremental Na■ve Bayesian Algorithms with Missing Data

在具有缺失数据的数据集的分类过程中,缺失的数据中蕴含着有用的信息未被考虑的情况会引起分类精度的下降。增量式的学习能够利用不断加入的信息更新学习模型,并充分利用先验信息求解当前问题。给出了一个利用朴素贝叶斯分类模型实现对缺失数据的增量分类的算法。该算法在增量学习的过程中考虑了缺失数据和先验信息对分类器的影响。

The precision of classifier with missing data will decrease if the useful information in the missing data is not be considered. Incremental learning can update study model based on the adding new data, and the prior information is fully used to solve current problem. The incremental Naive Bayesian learning algorithms with missing data is provided. The algorithms consider the effect of missing data and prior knowledge to the precision of classifier during incremental learning.