基于LSTM的文本情感分析方法

Text Sentiment Analysis Method Based on LSTM

摘要: 随着互联网的发展和电子商务的兴起，人们通过各种社交、电商平台发表自己的看法与见解，从这些用户评论数据中准确地挖掘出有用的信息是当前的研究热点。针对网络中的各类文本评论数据，本文基于深度学习的方法对这些数据进行情感分析，采用长短期记忆(Long Short-Term Memory, LSTM)神经网络模型构建情感分类器，对文本的情感倾向进行预测与分类。实验表明基于LSTM的情感分析方法可以很好地解决长距离依赖问题，具有较好的分类效果。

Abstract: With the development of the Internet and the rise of e-commerce, people express their views and opinions through various social and e-commerce platforms. It is a challenging issue to accurately mine useful information from user comment data. For all kinds of text comment data in the network, sentiment of the comment data is analyzed in this paper through deep learning. The Long Short-Term Memory (LSTM) neural network model is used to construct a sentiment classifier to predict and classify the sentiment tendency of the text. The experimental results show that the sentiment analysis method based on LSTM can solve the problem of long-distance dependence and has a good classification effect.