MapReduce Based Implementation of Tree Augmented Navive Bayes Algorithm
CHEN Yan;CHEN Ya-lin;LAN Si-mei;School of Mathematics and Information Science of Guiyang University;School of Management Science,Nanjing University of Finance & Economics;
To copy the contradiction between shortening the classifying time and increasing data in big data inverionment,as well as low efficiency and high complexity of application on serial Bayesian classifier,the MapReduce based implementation of tree augmented Navive bayes algorithm is proposed. In the algorithm,to enhance the classification accuracy,tree augmented naive Bayes algorithm,which weakens the data independence,is employed. Meanwhile,to lower the responding time of the algorithm,MapReduce model is introduced into the algorithm,which changes the algorithm from serial to parallel. The experiment results show that,compared with the traditional algorithm,the algorithm proposed in this paper has higher efficiency,and better speedup can be generated with more data nodes.
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.