【摘要】 由于当前已有攻击识别方法未对网络实验室虚假数据进行预处理，导致计算开销、存储开销以及能量消耗均较高，攻击的识别率偏低。现提出一种网络实验室虚假数据注入攻击深度识别方法。分析网络实验室虚拟数据的注入机理，滤除虚假数据噪声。将决策树算法和梯度提升框架结合，构建虚假数据注入攻击深度识别模型。采用自适应混沌果蝇算法优化模型参数，将经过预处理的虚假数据输入到识别模型中，根据虚拟数据的更新门以及重置门，获取虚假数据注入攻击训练数据集，实现虚拟数据注入攻击的深度识别。仿真结果表明：所提方法可以有效提升识别率，降低计算开销、存储开销以及能量消耗。实验结果证明了所提方法具有较好的应用前景。更多还原

【Abstract】 Due to the lack of preprocessing for false data in network laboratories using existing attack recognition methods, the computational, storage, and energy consumption are all high, resulting in a low recognition rate for attacks. A deep identification method for false data injection attacks in network laboratories is proposed. At first, the injection mechanism of virtual data in network laboratory was analyzed, and then false data noise was filtered out. After that, the decision tree algorithm was combined with gradient boosting framework to construct a model of deep recognition for false data injection attack. Moreover, the adaptive chaotic fruit fly optimization algorithm was used to optimize the parameters of the model, and then the preprocessed false data was input into the model. Furthermore, the training data set of false data injection attacks was obtained by the update gate and reset gate of data. Finally, the deep recognition of false data injection attack was achieved. Simulation results show that the proposed method can effectively improve the recognition rate and reduce the computational overhead, storage overhead and energy consumption, which proves that the proposed method has good application prospects.更多还原