基于循环神经网络的互联网短文本情感要素抽取

Sentiment Classification of the Short Texts on Internet Based on Convolutional Neural Networks

针对大数据互联网短文本信息,比较几种深度循环神经网络(Recurrent Neural Networks,RNN)模型,提出了一种基于双向长短时记忆(Bidirectional Long Short-Term Memory,BLSTM)的循环神经网络模型的互联网短文本情感要素抽取方法。实验结果表明,该方法不仅可以有效完成互联网短文本中情感要素抽取工作,而且明显提高了抽取准确率。

For the information of big data Internet essay,a deep convolution neural network(convolutional neural networks,CNNs)model of the short text on the Internet is put forward.And first use the Skipgram in the Word2 vec training model of feature vector,then further extracting feature vector into CNNs,finally training the classification model of the depth convolution neural network.The experimental results show that,compared with classification methods of traditional machine learning,this method not only could effectively handle Internet emotion classification in this essay,but also significantly improves the accuracy of emotion classification.