基于LSTM的复杂网络文本情感分析模型构建

Construction of Complex Network Text Sentiment Analysis Model Based on Deep Learning

互联网上存在海量的网络文本数据,如何对这些文本进行有效情感分析是文本数据挖掘面临的一个挑战。情感分析可以对大量文本数据情感极进行性分类,得出正面、负面或中性的文本情感表达,对于信息监管和舆情分析都有很重要的应用价值。由于网络文本的语义、语法结构复杂,一般模型对于文本的语义理解较困难,因此本文提出了基于长短期记忆(LongShort-TermMemory,LSTM)的深度学习情感分析模型。将从语义提取、深度语义嵌入等角度,构建融合语义知识的深度学习情感分析模型,实现网络文本的有效情感分类。

Emotion analysis can classify the emotional polarity of a large number of text data and obtain positive,negative or neutral text emotional expression,which has very important application value for information supervision and public opinion analysis.Due to the complex semantic and grammatical structure of Web text,it is difficult for general models to understand the semantics of text.Therefore,this paper proposes a deep learning emotion analysis model based on long short-term memory(LSTM).From the perspective of semantic extraction and deep semantic embedding,a deep learning emotion analysis model integrating semantic knowledge will be constructed to realize the effective emotion classification of Web text.