基于BGSA和ELM的垃圾邮件检测  

The Spam Detection Based on BGSA(Binary Gravitational Search Algorithm) and ELM(Extreme Learning Machines)

作　　者：李斌[1] 韩晓红[2] 

机构地区：[1]国网宁夏电力公司信息通信公司,宁夏银川750001 [2]太原理工大学新型传感器与智能控制教育部重点实验室,山西太原030024

出　　处：《山西科技》2015年第4期130-132,共3页Shanxi Science and Technology

基　　金：山西省自然科学基金资助项目(项目编号:2014011021-1)

摘　　要：提出了一种新的垃圾邮件检测方法。该方法采用二进制万有引力搜索算法(Binary Gravitational Search Algorithm,BGSA)提取关键特征,使用极限学习机(Extreme Learning Machines,ELM)作为分类器进行垃圾邮件分类。收集了2014年6 000封电子邮件,比较BGSA与其他启发式优化算法,结果表明BGSA的分类性能优于GA(Genetic Algorithm)和BPOA(Binary Particle Optimization Algorithm)。This paper puts forward a new method for spam detection, which adopts BGSA to extract the key features and adopts ELM as the classifier to carry out the spam classification, collects 6 000 E-mails in the year of 2004, and makes the comparison of BGSA with other heuristic optimization algorithms, and based on the results, points out that the classification capability of BGSA is better than that of GA and BPSO.

关 键 词：垃圾邮件检测 特征提取 GSA算法 ELM算法 

分 类 号：TP393.07[自动化与计算机技术—计算机应用技术]