基于遗传算法和随机森林的入侵检测方法研究

INTRUSION DETECTION METHOD BASED ON GENETIC ALGORITHMAND RANDOM FOREST

入侵检测系统中,待测数据通常存在特征数量多、具有冗余性和相关性的特点,导致检测准确率降低、检测时间增加。提出一种基于多层感知机的遗传算法,建立4层感知机神经网络,将网络的分类能力作为遗传算法适应度评价方法,筛选出最优特征子集,建立随机森林分类器,使用网格验证方法确定随机森林超参数值,利用选取出的特征子集进行入侵类型识别。实验结果表明,该方法在KDD99数据集上对正常和22种类别的入侵数据平均检测准确率达到92%以上,并且具有较好的实时性。

In intrusion detection system,the number of original features is large,and there are redundant features and related features,whichreduces the detection accuracy and increase the detection time.A genetic algorithm based on multi-layer perceptron was proposed.A four-layer perceptron neural network was established and the classification ability of the network was used as the fitness evaluation method of genetic algorithm to select the optimal feature subset.The random forest classifier was established,and the value of the hyper parameters were determined by grid verification method.The random forest classifier used the optimal feature subset to identify intrusion types.The experimental results show that the average detection accuracy of normal and 22 types of intrusion data is more than 92%on KDD99 data set,and with a good real-time performance.