计算机工程与应用 ›› 2022, Vol. 58 ›› Issue (17): 13-22.DOI: 10.3778/j.issn.1002-8331.2111-0219

• 热点与综述 • 上一篇    下一篇

俄语情感分析研究综述

徐琳宏,刘鑫,阎月,原伟,林鸿飞   

  1. 1.大连外国语大学 语言智能研究中心,辽宁 大连 116044
    2.锦州师范高等专科学校,辽宁 锦州 121000
    3.信息工程大学 洛阳校区,河南 洛阳 471003
    4.大连理工大学 计算机系,辽宁 大连 116024
  • 出版日期:2022-09-01 发布日期:2022-09-01

Survey of Russian Sentiment Analysis

XU Linhong, LIU Xin, YAN Yue, YUAN Wei, LIN Hongfei   

  1. 1.Research Center for Language Intelligence, Dalian University of Foreign Languages, Dalian, Liaoning 116044, China
    2.Jinzhou Normal College, Jinzhou, Liaoning 121000, China
    3.Information Engineering University, Luoyang, Henan 471003, China
    4.Department of Computer, Dalian University of Technology, Dalian, Liaoning 116024, China
  • Online:2022-09-01 Published:2022-09-01

摘要: 社交媒体中蕴含着用户的大量观点和评论,从中提取情感信息,有助于了解俄语区民众对热点事件、产品和服务等的真实态度,为相关政策的制定和调整提供依据,进而促进区域内国家间的合作共赢。按情感分析的流程从资源建设和自动识别两个方面详细梳理了俄语情感分析领域的研究现状,并在此基础上对比分析了各类方法在不同数据集上性能和特征选择方案。研究结果发现俄语语料等资源的数据来源需要拓宽,且同类资源还可以进一步整合,自动识别方面主流的识别模型为机器学习和深度学习两种,整体识别准确率还有待提高。通过综述该领域的不足,探索了未来可能的研究方法,为进一步研究提供借鉴。

关键词: 俄文, 情感分析, 情感资源, 自动分类

Abstract: Social media contains plenty of users’ views and comments. Extracting emotional information from social media helps to understand the real attitude of people in Russian speaking areas towards hot events, products and services, provide basis for the formulation and adjustment of relevant policies, and then promote win-win cooperation among countries in the region. According to the process of emotion analysis, the status quo in the field of Russian emotion analysis from two aspects of resource construction and automatic recognition is introduced; the performance and feature selection schemes of various methods on different data sets are analyzed. The results show that the data sources of Russian corpus and other resources need to be broadened, and similar resources can be further integrated. The mainstream recognition models in automatic recognition are machine learning and deep learning, and the overall recognition accuracy needs to be improved. By summarizing the shortcomings in this field, the possible research methods in the future are explored, and the references for further research are introduced.

Key words: Russian, sentiment analysis, sentiment resource, automatic classification