Website classification method and device

Abstract

Embodiments of the invention provide a website classification method and device. The method comprises the following steps of: obtaining a to-be-classified website; crawling a page text and a keyword in the to-be-classified website; calculating an occurrence frequency, in the keyword, of each preset website category label so as to obtain a first classification result set, wherein each website category label comprises a label name and synonyms thereof; inputting the page text and the keyword into a pre-configured Bayesian classification model so as to obtain a second classification result set, wherein the second classification result set comprises a predicted probability value of each website category label, and training samples of the Bayesian classification model are obtained through website crawling; and outputting a classification result on the basis of the first classification result set and the second classification result set. By adoption of the method and device, manual processing is not needed, the workload is small, and correct website classification can be realized when webpage text data is less.

Description

technical field [0001] The present application relates to the field of computer technology, in particular, to a website classification method and device. Background technique [0002] The website classification methods in the prior art mainly generate training samples by manual labeling, extract features based on web page content, and use machine learning algorithms to learn training samples to obtain a Bayesian classification model, thereby realizing website classification. However, the above method requires manual labeling of training samples, which is a huge workload, and if the content of the webpage is small, the error of the classification result of the machine learning algorithm is relatively large. [0003] In addition, there is also a method of crawling a large number of websites, generating training samples through clustering algorithms and manual annotation, thereby extracting features based on web page content and using machine learning algorithms to learn traini...

AI Technical Summary

Problems solved by technology

However, the categories generated by the clustering algorithm are relatively coarse, which may not meet the actual needs. Similarly, if the content of the web page is small, the classification result of the machine learning algorithm is very poor.

Images