摘要
社交网络用户生成的内容具有多样性和非结构化的特点,导致文本数据中蕴含的情感复杂多变,难以准确识别与分类.为此,提出基于多因子权重与GloVe模型的社交网络用户情感主题分类方法.利用数据挖掘技术采集社交网络用户的电子文本,提取其中的多因子权重,结合GloVe模型分析文本与主题之间的关系,从而对文本语义进行增强,引入核主成分分析方法提取并选择最有效的文本分类特征,以此为依据,以文本特征作为支持向量机分类器的输入,从而根据待测文本的类别概率确定文本的情感类型.实验结果表明,利用所提方法对不同数据集进行情感主题分类,得到的对数损失率始终保持在0.40%以内,整体分类精度较高.
The content generated by social network users has diverse and unstructured characteristics,which leads to complex and varied emotions contained in text data,making it difficult to accurately identify and classify.Therefore,a social network users sentiment topic classification method based on multi factor weights and GloVe model is proposed.Using data mining techniques to collect electronic texts from social network users,extracting multiple factor weights,and combining GloVe model to analyze the relationship between text and topic,thereby enhancing text semantics.Introducing kernel principal component analysis method to extract and select the most effective text classification features,which are used as input for support vector machine classifier,to determine the sentiment type of the text according to the category probability of the tested text.The experimental results show that using the proposed method for sentiment topic classification on different datasets consistently maintains a logarithmic loss rate within 0.40%,indicating high overall classification accuracy.
作者
席文
XI Wen(School of Big Data,Fuzhou University of International Studies and Trade,Fuzhou 350001,China)
出处
《西安文理学院学报(自然科学版)》
2025年第2期35-42,共8页
Journal of Xi’an University(Natural Science Edition)
关键词
数据挖掘
社交网络
用户
情感主题
文本分类
data mining
social networks
users
emotional themes
text classification
