A URL-based webpage classifier construction method and its classification method

Abstract

The invention discloses a method for constructing a web page classifier based on a URL and a classification method thereof. Firstly, URLs of a plurality of web pages are obtained, web page attributesare marked for each URL, and each URL marked with the web page attributes is used as a training sample to form a training sample set. For each training sample in the training sample set, each trainingsample is segmented by selected characters, and then converted into a word vector. The word vector of each training sample marked with the attributes of the web page in the training sample set is used as input to train the convolution neural network, and the web page classifier is obtained. For the web pages that need to be classified, firstly, the URL of the web page is obtained as a test sample; then the selected characters are processed by word segmentation, and finally converted into word vectors. The word vector of the test sample is inputted into the web page classifier, and the classification result is outputted through the web page classifier. The invention greatly improves the classification accuracy rate of the malicious webpage.

Description

technical field [0001] The invention relates to the technical field of information security, in particular to a method for constructing a URL-based web page classifier and a classifying method thereof. Background technique [0002] The openness and virtualization of the Internet pose serious challenges to privacy, data and transaction security. In recent years, the use of malicious webpages for cybercrime has been rampant. According to statistics, nearly one-third of web pages are potentially malicious. Malicious web pages attack users by sending spam, phishing, etc., causing users without any security awareness to suffer various damages, including Loss of funds, theft of private information, etc., seriously threaten the security of users' property and information. For this reason, how to identify malicious webpages in a timely and effective manner has become an important problem to be solved urgently. [0003] In the prior art, it is generally recognized whether a webpage...

AI Technical Summary

Problems solved by technology

[0003] In the prior art, it is generally recognized whether a webpage is a malicious webpage by detecting the content and behavior of the webpage; when identifying a malicious webpage by detecting the content of the webpage, it is necessary to detect the text and picture content of the webpage, malicious code fragments, Behavior records in logs, etc. Therefore, identifying malicious webpages through the content of malicious webpages cannot avoid problems such as changing webpage content, encryption or equivalent replacement

Images