深度学习技术在高校网络舆情分析中的应用

Application of Deep Learning Technology in University Network Public Opinion Analysis

随着社交媒体在高校学生生活中的不断普及,网络舆情的管理变得日益关键。传统机器学习的自然语言处理存在对人工标记特征的强依赖、耗时和维度爆炸等问题。本文应用基于CNN的深度学习技术分析高校网络舆情。通过收集百度贴吧、小红书等社交平台的校园热点话题,运用Word2vec模型生成词向量,并基于CNN提取特征进行情感倾向分类。实验结果显示,基于CNN的情感分类准确率为90.82%,比传统的支持向量机高4.54%,比K-近邻法高2.17%。本文方法可为实现高校治理现代化提供一定的实践支持。

With the continuous popularization of social media in the lives of college students,the management of online public opinion has become increasingly crucial.The natural language processing of traditional machine learning has problems such as strong dependence on manually labeled features,time consumption,and dimensional explosion.This article applies CNN based deep learning technology to analyze online public opinion in universities.By collecting campus hot topics from social platforms such as Baidu Tieba and Xiaohongshu,Word2vec model is used to generate word vectors,and sentiment tendency classification is performed based on CNN feature extraction.The experimental results show that the accuracy of sentiment classification based on CNN is 90.82%,which is 4.54%higher than traditional support vector machines and 2.17%higher than K-nearest neighbor methods.This article's method can provide practical support for achieving modernization of university governance.