关键词: 朴素贝叶斯, 并行化, 半监督, 文本分类, 海量数据

Abstract: Nowadays TBs or even PBs data burst out every day, but there are so few labeled instances for training. For these two problems, this paper combines a semi-supervised Naïve Bayes algorithm and the Map-Reduce programming model, and proposes a new algorithm called parallelized semi-supervised Naïve Bayes (PSNB) algorithm. Experimental results show that the proposed algorithm can tackle with massive data efficiently, and use the unlabeled instances to improve the performance of the classifier.

Key words: Naïve Bayes, parallelization, semi-supervised, text classification, massive data