GRU神经网络下电力辅助决策模型疑点数据检测

Doubtful data detection of power aided decision-making model based on GRU neural network

为了保证电力辅助决策模型在电网业务中的应用价值,提出GRU神经网络下电力辅助决策模型疑点数据检测方法。模拟电力辅助决策模型的运行过程,获取模型运行数据。利用GRU神经网络提取数据特征,并运用随机森林算法确定当前数据类型。通过对疑点数据进行审计,确定电力辅助决策模型的疑点数据量,从而实现模型疑点数据的检测。实验结果表明,与传统方法相比,优化设计方法的疑点数据检测误差更小,表明该方法在检测性能方面具有明显优势。

In order to ensure the application value of power aided decision model in power grid business,a method of suspicious data detection of power aided decision model based on GRU neural network is proposed.The operation process of the power-assisted decision model is simulated and the running data of the model is obtained.GRU neural network is used to extract data features and random forest algorithm is used to determine the current data type.By auditing the doubtful data,the quantity of doubtful data of the power aided decision model is determined,so that the doubtful data of the model can be detected.The experimental results show that,compared with the traditional method,the optimization design method has less error in the detection of doubtful data,indicating that the method has obvious advantages in the detection performance.