基于深度学习的网络文本多粒度情感提取

Multi-granularity Sentiment Extraction of Web Text Based on Deep Learning

网络文本具有复杂的语义结构,提高模型的语义理解能力是当前情感提取模型面临的挑战。在此背景下,笔者提出了基于深度学习的多粒度融合情感分类模型,将语义先验知识引入到LSTM深度学习模型,提升模型对于情感语义的识别度,从而有效提高大规模文本数据情感分类的准确率。

Web text has a complex semantic structure,and improving the model’s semantic understanding is a challenge faced by the sentiment extraction.A multi-granularity fusion emotion classification model based on deep learning is proposed,which introduces semantic prior knowledge into the LSTM,improves the model’s recognition of emotion semantics,and effectively increase the accuracy of sentiment classification for large-scale text data.