计算机科学 ›› 2010, Vol. 37 ›› Issue (7): 11-15.

• 综述 • 上一篇    下一篇

情感分析与认知

李维杰   

  1. (中国科学院自动化研究所复杂系统与智能科学实验室 北京100190)
  • 出版日期:2018-12-01 发布日期:2018-12-01
  • 基金资助:
    本文受973国家重点基础研究发展计划(2007CB311007)资助。

Introduce Sentiment Analysis to Cognitive Science

LI Wei-jie   

  • Online:2018-12-01 Published:2018-12-01

摘要: 分析了情感分析的3个主要步骤,包括文本情感获取与表达、文本情感分类与计算以及文本情感分析的应用。情感分析得到的结论主要是对相关观点的摘要、对相关事件态度的预测或者统计等,但这些结论都没有发挥文本情感在认知中的作用。为了将情感分析应用于认知科学,提出了情感由情感信号和情感实体组成的观点。情感信号主要是指情感的一些形式载体,比如心跳加速、脸红等这些人体内外的某些表现,表达情感的文字、图片、声音等这类媒体。情感实体主要是指人类对情感形成的一种共识,比如爱、恨、僧恶、高兴、羞愧、嫉妒、内疚、恐惧、焦虑等与人的意识相关联的部分。同时提出了在人工智能中利用情感信息的设想。这对于模拟情感对认知的影响具有一定的意义。

关键词: 情感分析,观点挖掘,认知科学,思维科学,人工智能

Abstract: We concluded the process of sentiment analysis, which has three main steps, acquisition and expression of sentiment, classification and estimation of sentiment, and application of sentiment analysis. The results of sentiment analysis can be classified into summarization of textual sentiment, estimation or analysis sentiment to persons or events in texts. But nobody gives any attention to the role of emotion in texts as it in cognitive. For replying to this problem, we tried to introduce sentiment analysis to cognitive science and get emotion in texts by utilizing sentiment analysis. We proposed that emotion should have two constituent parts:emotion signal and emotion element. Emotion signals are the carriers of emotion. The carriers include phenomena observed in body such as face color, heart rate, and media which expresses or pass emotion such as images,words,voices. Emotion elements are the ones which are accepted by the society,such as love, hate, delight, abashment, envy, guilt, dread, anxiety and so on, being related to consciousness. We also proposed some ideas on introducing emotion to artificial intelligence. It may be helpful for cognitive science to analyze and utilize emotion.

Key words: Sentiment analysis, Opinion mining, Cognitive science, Noctic science, Artificial intelligence

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!