基于长短时记忆网络的电力故障维修效果情感分析

Sentiment analysis of power fault system maintenance effect based on Long Short-Term Memory

基于电力行业的服务特性,需要对用户各种自然语言评论和主观体验等进行情感分析,这样可以改进服务,提高服务质量,提高用户满意度.从电力故障维修数据中,利用基于深度学习的自然语言处理方法对维修结果记录进行情感分析,针对循环神经网络记忆性短的缺点,采用长短时记忆网络进行处理,克服了现在浅层学习不能自主提取特征、自主抽象,因而处理复杂事物能力有限,泛化能力差的缺点.本文在90000多条电力故障维修记录语料上进行了实验,与循环神经网络、卷积神经网络、径向基函数神经网络进行对比,证明本文基于长短时记忆网络的情感分析模型准确率更高,效果更好.

Based on theservice characteristics of the power industry,it is necessary to conduct sentiment analysis of various natural language comments and subjective experiences to improve service quality and improve customer satisfaction.In this paper,the deep learning-based natural language processing method is used to analyze the emotion of the maintenance result record from the power system fault maintenance data.Aiming at the shortcoming of short memory of Recurrent Neural Network(RNN),Long Short-Term Memory(LSTM)neural network with long-term and short-term memory is used for processing.It overcomes the shortcomings of shallow learning that cannot extract features and abstract independently,so it has limited ability to deal with complex things and poor generalization ability.The experiments are carried out on more than 90000 corpuses of power system fault maintenance records,and compared with RNN,Convolution Neural Network(CNN)and Radicdl Basis Function neural network.It proves that the model based on LSTM has higher accuracy and better effect.