Text sentiment analysis based on long short-term memory
暂无分享,去创建一个
With the rapid development of Internet and big explosion of text data, it has been a very significant research subject to extract valuable information from text ocean. To realize multi-classification for text sentiment, this paper promotes a RNN language model based on Long Short Term Memory (LSTM), which can get complete sequence information effectively. Compared with the traditional RNN language model, LSTM is better in analyzing emotion of long sentences. And as a language model, LSTM is applied to achieve multi-classification for text emotional attributes. So though training different emotion models, we can know which emotion the sentence belongs to by using these emotion models. And numerical experiments show that it can produce better accuracy rate and recall rate than the conventional RNN.
[1] Alex Graves,et al. Supervised Sequence Labelling with Recurrent Neural Networks , 2012, Studies in Computational Intelligence.
[2] NeyHermann,et al. From feedforward to recurrent LSTM neural networks for language modeling , 2015 .
[3] Ludek Müller,et al. Application of LSTM Neural Networks in Language Modelling , 2013, TSD.
[4] Hermann Ney,et al. Lattice decoding and rescoring with long-Span neural network language models , 2014, INTERSPEECH.