Research on Automatic Classification,Extraction and Analysis Techniques for Electric Power Engineering Data
To solve the problems of complex data types,large scale and difficult analysis and processing in the current electric power engineering data processing process,an automatic classification,extraction and analysis techniques for electric power egineering data based on the improved K-means clustering algorithm and the long short-term memory(LSTM)neural network is proposed.Aiming at the defects of the traditional K-means clustering algorithm,such as the strong uncertainty of K value selection,and the weak correlation of spatial distance characteristics between electric power data and cluster centroids,a K value selection and attribute weighting method based on the threshold decision is adopted to improve the spatial distance algorithm.Meanwhile,the LSTM is optimized by the sparrow search algorithm(SSA),which solves the problem of insufficient prediction accuracy due to the existence of uncertainty when the traditional network performs hyper-parameter selection.The results of tests on public power engineering datasets show that the accuracy and fault tolerance of the techniques classification are 92%and 98%,respectively,and the prediction accuracy of the cost data is more satisfactory.The research contributes to the analysis and prediction of engineering data.
Power engineeringAutomated analysisK-means clustering algorithmData classificationLong and short-term memory(LSTM)Sparrow search algorithm(SSA)
NETL
NSTL
万方数据
维普
