Research on Key Technologies of Public Opinion Analysis of Power Operation and Maintenance Based on Deep Learning

With the rapid development of network and social media, the operation and maintenance department of power supply enterprises is facing a complex situation of network public opinion. How to use deep learning method to analyze the public opinion information of power operation and maintenance is a topic worthy of our consideration. In view of the high cost of manual annotation of public opinion training in the field of power operation and maintenance, and combined with the characteristics of public opinion analysis in the field of power operation and maintenance, this paper adopts the active learning strategy and the dual channel convolution neural network (WEP-CNN) model which integrates the emotional polarity of words to analyze public opinion. The model greatly improves the labeling efficiency of data sets, and achieves significant improvement in accuracy, recall, F1 value and other indicators.