摘要
互联网的发展导致恶意软件信息类型种类繁多,为检测获取更深层次的恶意软件信息,研究基于知识图谱的恶意软件信息检测方法,提升恶意软件信息检测效果。利用文本挖掘技术中的Python网络爬虫技术,采集软件有效信息;通过信息增益算法,在采集的软件有效信息内,提取软件信息特征;在双向长短期记忆神经网络内输入软件信息特征,输出软件信息实体识别结果,并抽取软件信息实体间的关系;依据实体消岐技术,对抽取的软件信息实体关系进行知识融合,得到软件信息知识图谱;利用图推理算法,处理软件信息知识图谱,得到恶意软件信息检测结果。实验证明:该方法可有效采集软件有效信息,并提取软件信息特征,建立软件信息知识图谱;该方法可有效检测恶意软件信息,且检测精度较高。
The development of the Internet had led to a wide variety of types of malware information.In order to detect and obtain deeper levels of malware information,research was being conducted on malware information detection methods based on knowledge graphs to improve the effectiveness of malware information detection.Using Python web crawler technology in text mining to collect effective software information;Extract software information features from the collected software effective information through information gain algorithm;Input software information features into a bidirectional long short-term memory neural network,output software information entity recognition results,and extract relationships between software information entities;Based on entity disambiguation technology,knowledge fusion was performed on the extracted software information entity relationships to obtain a software information knowledge graph;Using graph inference algorithms to process software information knowledge graphs and obtain malware information detection results.Experimental results had shown that this method could effectively collect effective software information,extract software information features,and establish a software information knowledge graph;This method could effectively detect malicious software information and has high detection accuracy.
作者
桑道松
SANG Daosong(Huangshan Open University,AnHui Huangshan 245000)
出处
《九江学院学报（自然科学版）》
CAS
2024年第3期79-84,共6页
Journal of Jiujiang University：Natural Science Edition
基金
安徽省高等学校省级质量工程项目教学研究项目(编号2021jyxm0309)的研究成果之一。
关键词
知识图谱
恶意软件
信息检测
Python网络爬虫
神经网络
图推理算法
knowledge graph
malicious software
information detection
python web crawler
neural networks
graph inference algorithm
