基于深度学习的Wi-Fi网络入侵检测方法

Wi-Fi intrusion detection method based on deep learning

提出一种基于深度信念网络(DBN)的Wi-Fi网络入侵检测模型。使用SMOTE算法对数据样本中的异常数据和正常数据进行数量平衡,使用降噪自编码网络(DAE)对平衡后的数据进行降维,消除无关或冗余特征降低检测建模规模。在AWID数据集上进行实验,实验结果表明,与其它基于浅层学习算法的检测方法相比,所提方法可有效精简数据特征,降低检测时间,在检测精度和误报率方面体现出了更优性能。

A Wi-Fi network intrusion detection model based on deep belief network(DBN)was proposed.SMOTE was used to balance the data samples to improve the the classification detection performance.To reduce the dimensionality and redundancy of the feature vectors,denosining autoencoder network(DAE)was used to reduce the dimension of the feature vectors.Results of experiments on the AWID dataset show that compared with other shallow learning algorithms,the proposed model can effectively simplify data features and reduce detection time and false alarm rate,and it also shows better performance in detection accuracy.