Email type judgement method and device and establishment device of system and behavior model

Abstract

The invention discloses a judging method, device and system of mail pattern and model building device in the internet technical domain, wherein the judging method of mail pattern comprises the following steps: reading and classifying unknown mail head; extracting the field one satisfying the preset condition one from the mail head; vectorizing the field one and its display pattern composition; obtaining the characteristic vector one of the preset quantity one; inputting based on the characteristic vector one; prebuilding the reserved model data as line; using preset predicting algorism to calculate; getting the calculating result; classifying the unknown mail pattern to judge according to the calculated result. The invention also provides the corresponding device and system corresponding to the method, which improves the judging speed of the mail pattern.

Description

technical field [0001] The invention relates to Internet technology, in particular to a mail type judging method, device and system, and a behavior model building device. Background technique [0002] E-mail, as the largest application of the Internet, has always been favored by the majority of Internet users. However, in recent years, the spam problem has become increasingly serious. The basic feature of spam is "uninvited", and most spam has commercial or other publicity purposes. At the same time, the judgment of spam has a lot to do with the receiver of the mail, and different users may have different judgment results for the same mail. With the advancement of technology, spam filtering technology is changing from a single based on static rules and statistical classification to behavior-based filtering technology. [0003] The existing mainstream spam filtering methods are all based on email content. One spam filtering method is based on Learning Vector Quantization (...

AI Technical Summary

Problems solved by technology

[0007] The content of the email body is large and varied, which will lead to problems such as slow training and incomplete training set, which may lead to low filtering accuracy; and, because the content and format of the email body are uncertain, it may cause email judgment The speed is slow; further, the email experience of non-Chinese emails and other emails is expressed as a zero vector, so that the email is considered as a normal email, so when spam is also expressed as a zero vector, it cannot be filtered, which further reduces the efficiency of filtering. Correct rate

Images