Multi-Relational Naive B ayesian Classifier towards Relational Individual
The“Graph-NBalgorithm is the first try which integrates Naive Bayesian classifiers with relational database technology in the multi.relational area.Compared with methods based on ILP.Graph-NB is more efficient.However.Graph.NB without pruning has poor accuracy because of statistical bias.Tosolve the problem.this paper proposes a multi.relational Naive Bayesian classifier towards relational individuaI (RI.MRNBC).Its counting method is based on relational individual instead of tuples in current table. Accordingly. RI-MRNBC gives improved Naive Bayesian formula towards relationalindividual.Experiments show RJ-MRNBC can get high efficiency and good accuracy.
Multi-relational data mining naive bayes classification relational individuaI
Guangmei Xu Bingru Yang Bolin Wu Wei Song Yiqing Qin
国际会议
The International Conference Information Computing and Automation(2007国际信息计算与自动化会议)
成都
英文
432-434
2007-12-19（万方平台首次上网日期，不代表论文的发表时间）