The Data Cleaning of Electric Industrial Control Terminal Based on the iForest and Genetic BP Neural Network Algorithms

This paper focuses on the data cleaning of the electric industrial control terminal. First, we introduce the isolation forest algorithm for locating the dirty data. In order to improve the detection accuracy for higher dependent data in power system, the isolation forest is modified. Second, we use the BP neural network to predict the dirty data. To optimize the performance, the genetic algorithm is introduced for optimizing the weight of BP neural network. Finally, we provide a case study for illustrating the effectiveness of the proposed method.