基于改进樽海鞘群算法的垃圾邮件分类

An Improved Salp Swarm Algorithm for Spam Classification

为了规避电子邮件中的垃圾信息,提出一种基于改进樽海鞘群算法的垃圾邮件分类。因樽海鞘群算法缺少惯性参数和找到全局搜索潜在解决方案的能力,故利用Tent映射对初始种群施加混沌扰动,并在位置更新中加入了惯性权重策略。采用增强的算法优化分类器的参数,使得分类效果愈加显著。基于不同分类器和算法的实验表明,优化后的算法明显提高了垃圾邮件的分类精确度和最佳识别准确度。

In order to avoid spam in E-mail,an improved salp swarm algorithm for spam classification is proposed.However,the SSA lacks inertial parameters and ability to find global search potential solutions.The chaotic perturbation is applied to the initial population using Tent mapping,and the inertial weight strategy is added to the position update.Then,the parameters of the classifier are optimized by the enhanced algorithm,which makes the classification effect more significant.Finally,based on different classifiers and algorithm,experiments demonstrate that the optimized algorithm has an enormously increase on classification accuracy and the optimal recognition precision of spam.