Defect Text Analysis Method of Electric Power Equipment Based on Double-Layer Bidirectional LSTM Model

摘要

How to effectively deal with a large number of defect texts accumulated in the power system is one of the difficulties in the field of Chinese text classification technology. And the accuracy of the current power system defect text classification model needs to be further improved. In view of the above background, first of all, for the shortcomings of the Long Short-Term Memory (LSTM), the Deep Attention Mechanism is integrated. An optimized Bidirectional Long Short-Term Memory model based on deep attention mechanism is constructed. Then, using the power transformer defect text as the analysis object, the classification effect of DA-BiLSTM was tested. Finally, the classification effect is compared with several typical machine learning classification models. The experimental results of the simulation of the example show that, the classification effect of the DA-BiLSTM has a better advantage than several typical machine learning model.