摘要
情绪分析在日益发展的互联网时代极为重要,其中的研究内容主要包括语料集的选择和自然语言处理技术。为了对网民的个人情绪有较全面的了解,本文提出一种在用户授权的情况下,通过抓取输入法本地词库中的输入内容,并使用BERT和LSTM模型分析用户情感状况和情感倾向的情绪分析方法。当用户心理健康状态有下滑趋势或已处于状态较差时,则会采取相应的干预措施,以达到关怀用户的效果。
Sentiment analysis is extremely important in the growing Internet era,where research mainly includes corpus selection and natural language processing techniques.In order to have a more comprehensive understanding of the personal emotions of netizens, this paper proposes an emotion analysis method by grabbing the input content in the local thesaurus of the input method,and using Bert and LSTM models to analyze the emotional status and emotional tendency of users under the condition of user authorization.When the user ’s mental health is on a downward trend or is already in a poor state,corresponding interventions will be taken to care for the user.
作者
任宇翔
程丽霏
赵天一
王溯
孙庚辉
刘芳
REN Yuxiang;CHENG Lifei;ZHAO Tianyi;WANG Su;SUN Genghui;LIU Fang(Medical Image College of Mudanjiang Medical College,Mudanjiang Heilongjiang 157011;Modern Educational Technology Center of Mudanjiang Medical College,Mudanjiang Heilongjiang 157011)
出处
《软件》
2022年第8期53-57,共5页
Software
基金
黑龙江省基本科研业务费项目(2020-KYYWF-0790)。
