基于神经网络的文本情感分类模型研究

Research on text sentiment analysis based on neural network

情感分类作为近年来自然语言处理领域的热门研究方向,旨在识别文本中的情感态度,如积极、消极或者中立等,对社交媒体、新闻、评论和用户反馈等大量文本数据进行挖掘并分析其情感极性对于研究者和政府来讲具有十分重要的意义。传统的情感分类算法通常使用基于统计的特征提取方法,如词袋模型,再结合机器学习算法,如支持向量机(SVM)和朴素贝叶斯分类器等来进行分类。在基于对神经网络的研究下实现文本情感分析,对文本数据集进行预处理操作后建立情感分析模型,使用Keras框架搭建循环神经网络以识别情感倾向,定义相关函数后进行模型训练,并采用一系列方法指标来进行模型评估检验模型性能,比较传统机器学习算法提高了情感分析的精度和效率。

As a popular research direction in the field of natural language processing in recent years,sentiment classification aims to identify the emotional tendency in text,such as positive,negative or neutral,etc.,and it is of great significance for research⁃ers and governments to mine and analyze the emotional polarity of a large amount of text data such as social media,news,comments and user feedback.Traditional sentiment classification algorithms usually use statistical⁃based feature extraction methods,such as bag⁃of⁃words models,combined with machine learning algorithms,such as support vector machines(SVMs)and naïve Bayes classi⁃fiers.Based on the research of neural network,the text sentiment analysis is realized,and the sentiment analysis model is estab⁃lished after the text data set is preprocessed,use the Keras framework to build a recurrent neural network to identify emotional ten⁃dencies,after defining the relevant functions,model training is performed,and use a series of methods to test the model perfor⁃mance.Compared with traditional machine learning algorithms,the accuracy and efficiency of sentiment analysis are improved.