Keyword semantic classification method and system for sensitive data leakage detection

Abstract

The invention discloses a keyword semantic classification method and system for sensitive data leakage detection. The keyword semantic classification method comprises the following specific steps: 1,inputting a sensitive keyword library; 2, vectorizing the keyword library by using a natural language processing technology; 3, carrying out dimensionality reduction on the vector data corresponding to each keyword; 4, performing clustering analysis on the vector data subjected to dimension reduction; 5, for the keyword vectors subjected to clustering analysis, identifying keyword categories in combination with services; 6, performing keyword grouping optimization on each category according to the category labels which have been identified; and 7, outputting the category of the sensitive keyword library. The vocabularies are expressed in a vectorization form to achieve classification of the vocabularies; vocabulary classification optimization is realized by using similarity calculation based on vocabularies of a specific category; and a huge keyword library is classified and refined, so that the working efficiency of a user and the data matching accuracy of a specified field are improved.

Description

technical field [0001] The invention relates to the technical field of computer data security, in particular to a keyword semantic classification method and system for sensitive data leakage detection. Background technique [0002] Sensitive data generally refers to information and data with a high degree of confidentiality of enterprises, organizations or individuals. [0003] With the rapid development of the Internet in recent years, information security has become particularly important, and the risk of sensitive data leakage faced by enterprises is also increasing. For such problems, the most widely used method is to use sensitive word matching method to discover Internet sensitive information, and the sensitive word lexicon used in it will become larger and more complicated as the sample data gradually increases. [0004] There are many types of sensitive data, and the extracted sensitive keywords have the characteristics of a wide range and variety. How to effectivel...

AI Technical Summary

Problems solved by technology

But there are still some important sensitive words missing

[0007] To sum up, the method of using keywords to match sensitive data in the prior art has disadvantages such as high computational resources occupation, long time period, and lack of pertinence in the matching field.

Images