计算机应用与软件2016，Vol.33Issue(8)：57-61,5.DOI:10.3969/j.issn.1000-386x.2016.08.013
基于 mRMR 的多关系朴素贝叶斯分类
MULTI-RELATIONAL NAIVE BAYESIAN CLASSIFICATION BASED ON MRMR
摘要
Abstract
In classification task,feature selection is an important method to improve classification effect.In real life,data is stored in multiple relational databases.There are many irrelevant and redundant features in multiple relational database,and they have little or even no contribution to classification task.How to effectively apply the feature selection to multi-relational classification is rather important. Therefore,we applied the feature selection method of maximum relevance minimum redundancy to multi-relation classification,the feature selection is carried out on every relation table in database and to pick out the feature sets with better effect on classification.Then,we used the multi-relational naive Bayesian classification algorithm to classifying and testing the multi-relational database with the features selected. Experimental results also showed that the performance of the algorithm has been improved.关键词
多关系/分类/特征选择Key words
Multi-relational/Classification/Feature selection分类
信息技术与安全科学引用本文复制引用
张晶,毕佳佳,刘炉..基于 mRMR 的多关系朴素贝叶斯分类[J].计算机应用与软件,2016,33(8):57-61,5.基金项目
国家自然科学基金项目（61273292，61305063）。 （）
