摘要
提出基于卷积神经网络(CNN)的电力工程造价数据异常识别方法。经过采集和查找大量数据,进行预处理和特征提取后,设计了一个包含卷积层、池化层和全连接层的CNN模型来学习和识别异常造价数据的模式。实验证明,该方法在准确性和鲁棒性方面表现优秀,具有潜在的应用前景。
This article proposes a method for identifying anomalies in power engineering cost data based on convolutional neural networks(CNN).After collecting and searching for a large amount of data,after preprocessing and feature extraction;a CNN model consisting of convolutional layer,pooling layer,and fully connected layer was designed to learn and identify patterns of abnormal cost data.Experiments have shown that this method performs well in terms of accuracy and robustness,it has potential application prospects.
作者
陈天宇
Chen Tianyu(Guangdong Power Grid Co.,Ltd.Shanwei Power Supply Bureau,Shanwei Guangdong 516600,China)
出处
《现代工业经济和信息化》
2023年第11期322-324,共3页
Modern Industrial Economy and Informationization
关键词
卷积神经网络
电力敏感数据
泄露风险监测
特征提取
convolutional neural network
electricity sensitive data
leakage risk monitoring
feature extraction
