摘要
为了提高评论情感分类与质量检测的准确率,提出了基于联合分类模型的评论情感分类与质量检测的综合性模型JNBERT,通过融合文本的情感与评论质量表示,经由softmax层获取了每一类的概率,并以概率最大一类作为情感分类和评论质量检测的最终结果,通过实验,证明本文提出的JNBERT模型能有效提高在线评论的情感分类和评论质量检测的效果,验证了情感分类与评论质量检测任务的相关性。
In order to improve the accuracy of review sentiment classification and quality detection,this paper proposes a comprehensive model JNBERT for review sentiment classification and quality detection based on a joint classification model,integrates the sentiment of the text and the quality of the review,and obtains the probabilities for each category through the softmax layer,and takes the highest probability category as the final results of sentiment classification and review quality detection,and through the experiments,proves that the JNBERT model can effectively improve the effect of sentiment classification and review quality detection of online reviews,and verifies the correlation between sentiment classification and review quality detection tasks.
出处
《图书情报导刊》
2022年第10期52-58,共7页
Journal of Library and Information Science
基金
苏州市图书馆学会2021年课题重点项目“人工智能技术实现图书馆智慧化管理与服务的研究”(项目编号:21-A-04)。