一种改进的朴素贝叶斯文本分类方法

朴素贝叶斯分类器是一种简单有效的文本分类方法.改进方法利用同义词对文本的特征词集进行过滤,在一定程度上放松了朴素贝叶斯的特征独立性假设;在特征选择时迭代了2种不同的特征选择方法,有效地提高了特征集的代表性.实验结果表明,本方法有效地提高了朴素贝叶斯分类器的性能.

An Improving Text Categorization Method of Na(i)ve Bayes

Nave Bayesian Classifier is a simple and powerful method for text classification.Our improved method uses synonyms to filtrate the features of text,loosing the independent condition required by Bayesian method.We apply two different feature extraction methods in the iteration process,enhancing the representative ability of feature collection effectively.The experimental results show that our advanced method has obviously improved the performance of Nave Bayesian Classifier.