| 蜣螂算法优化CNN-LSTM-AT模型的异常用电行为检测方法 |
 点此下载全文 |
| 引用本文：于多1，2，钱承山3，毛伟民3.蜣螂算法优化CNN-LSTM-AT模型的异常用电行为检测方法[J].计算技术与自动化,2025,(3):36-43 |
| 摘要点击次数: 391 |
| 全文下载次数: 220 |
|
|
| 中文摘要:针对用电模型预测精度不高导致异常检测效率低的问题，本文提出使用蜣螂优化算法（DBO）对卷积神经网络（CNN）-长短期记忆网络（LSTM）-注意力机制(AT)模型进行超参数选择的方法。首先，用户用电数据经过处理后，使用CNN对数据进行特征提取，用作LSTM的输入以分析时间序列，注意力机制提取LSTM隐藏状态的重要特征，忽略无用特征，逐步提高预测精度。其次利用DBO算法对CNN卷积层的特征检测器大小、LSTM网络的神经元大小和Dropout的大小进行优化，以提高模型性能，进而将预测结果与用户的用电数据进行比较和异常判断。最后，将本文所提方法在办公区域用电数据集上进行实验，实验验证了所提方法的有效性。 |
| 中文关键词:异常用电 蜣螂优化算法 注意力机制 卷积神经网络 长短期记忆网络 |
| |
| Abnormal Electrical Behavior Detection Method of DBO-CNN-LSTM Based on Attention Mechanism |
|
|
| Abstract:Aiming at the problem that the prediction accuracy of the electricity consumption model is not high, resulting in low anomaly detection efficiency, this paper proposes a method for hyperparameter selection of convolutional neural network (CNN)-long short-term memory network (LSTM)-attention mechanism (AT) model using dung beetle optimization algorithm (DBO). Firstly, after the user's electricity data is processed, the CNN extracts the features of the data, which is used as the input of the LSTM to analyze the time series, and the attention mechanism extracts the important features of the hidden state of the LSTM, ignores the useless features, and gradually improves the prediction accuracy. Secondly, the DBO algorithm is used to optimize the size of the feature detector of the CNN convolutional layer, the neuron size of the LSTM network and the size of Dropout, so as to improve the performance of the model, and then compare the prediction results with the user's power consumption data and make abnormal judgment. Finally, the proposed method is experimented on the electricity consumption dataset of the office area, and the effectiveness of the proposed method is verified by experiments. |
| keywords:abnormal electricity consumption dung beetle optimization algorithm attention mechanisms convolutional neural networks long short-term memory networks |
| 查看全文 查看/发表评论 下载pdf阅读器 |
|
|
|
位访问者,今日访问
