Design of Mechanical Information Text Classifier

A information text classifier of machinery-oriented is proposed in this paper.The method of document frequency and grey relation analysis are used to select feature, which reduce the characteristics of dimensionality, weaken relation between feature words and create the conditions for the Bayes.The Bayesian Classifier is ameliorated by using word's kinds-difference as weighted factor.The experimental results indicate that the classifier is able to improve recall and precision, and is useful in practice.