摘要
为准确识别意见领袖的敏感舆论倾向,有效把控敏感类舆情的发展,提出基于多任务学习的敏感舆论倾向识别模型(MTL-SA-LSTM)和基于指纹汇聚技术的快速识别模型。以准确识别意见领袖的敏感舆论倾向为目标,兼顾其识别效率。采用指纹汇聚技术关联原始敏感词和变形敏感词,采用语义指纹技术快速识别重复或相似度较高文本的敏感舆论倾向,通过MTL-SA-LSTM模型,对文本中的敏感舆论及舆论倾向两个任务进行识别。对比实验结果表明,该模型具有较高的识别准确率及识别效率。
To accurately identify the sensitive public opinion tendencies of opinion leaders,effectively control the development of sensitive public opinions,the sensitive public opinion tendency recognition model based on multi-task learning(MTL-SA-LSTM)and the fast recognition model based on fingerprint aggregation technology were proposed.The goal was to accurately identify the sensitive public opinion tendencies of opinion leaders,taking into account their identification efficiency.Fingerprint aggregation technology was used to associate the original sensitive words and deformed sensitive words.Semantic fingerprinting technology was used to quickly identify the sensitive public opinion tendency of repetitive or highly similar text.Through the MTL-SA-LSTM model,the sensitive public opinion in the text and public opinion tends were identified.Results of comparative experiments show that the proposed model has higher recognition accuracy and recognition efficiency.
作者
宋振
徐雅斌
SONG Zhen;XU Ya-bin(School of Computer,Beijing Information Science and Technology University,Beijing 100101,China;Beijing Key Laboratory of Internet Culture and Digital Dissemination Research,Beijing Information Science and Technology University,Beijing 100101,China)
出处
《计算机工程与设计》
北大核心
2021年第11期3293-3300,F0003,共9页
Computer Engineering and Design
基金
国家自然科学基金项目(61672101)
网络文化与数字传播北京市重点实验室基金项目(ICDDXN004)
信息网络安全公安部重点实验室开放课题基金项目(C18601)。
关键词
社交网络
敏感舆论倾向
多任务学习
指纹汇聚
语义指纹
social network
sensitive public opinion tendency
multi-task learning
fingerprint aggregation
semantic fingerprint
作者简介
宋振(1993-),男,河南商丘人,硕士研究生,研究方向为自然语言处理、大数据;徐雅斌(1962-),男,辽宁锦州人,硕士,教授,CCF高级会员,研究方向为云计算与大数据、社交网络、未来网络。E-mail:xyb@bistu.edu.cn。
