基于粗集的朴素贝叶斯分类算法及其应用

朴素贝叶斯方法是数据库分类知识挖掘领域一项基本技术,具有广泛的应用。论文针对朴素贝叶斯方法的限制,提出了基于粗集理论的贝叶斯的分类知识挖掘方法。该方法首先基于粗集理论的属性约简能力,根据数据库中条件属性和决策属性之间的依赖关系,进行属性的约简处理,然后基于朴素贝叶斯方法进行分类知识挖掘。实验结果表明,基于粗集理论的贝叶斯分类方法改善了贝叶斯分类方法中属性之间独立的限制,简化了挖掘模型,使挖掘性能具有明显的优化。

Rough set based na(i)ve bayesian classification algorithm and it's application

NaIve Bayesian method is basic technology that has been applied widely for class knowledge discover in database.Since the restriction of attribute dependency in naIve Bayesian method,Rough set theory based Bayesian class knowledge mining method has been proposed.With this method,dependency relation between conditional attributions and class attribution has been taken synthetically into account,attributes of conditional attributions has been reduced based on rough set with its ability of attribution reduction,finally naIve Bayesian method with reduced attributions has been used for the class knowledge discover in database.Experiment result indicates that rough set based Bayesian class model can ameliorate the restriction of attributions independence in naIve Bayesian method,simplify class mode mining model and optimize markedly performance of mining arithmetic.