Reflections on Sentiment/Opinion Analysis
Reflections on Sentiment/Opinion Analysis
复制标题
DOI：
10.1007/978-3-319-55394-8_3
复制
发表时间：
2017-01-01
期刊：
影响因子：
--
通讯作者：
Hovy, Eduard
中科院分区：
文献类型：
--
作者：
Li, Jiwei;Hovy, Eduard
在线文本中情感表达的检测已成为流行的自然语言处理应用。该任务通常被定义为识别给定文本片段中的单词或短语，其中读者理解作者表达了某人对某个主题的积极、消极或中立态度。这四个要素——表达词、态度持有者、主题和态度价值——在文献中几乎没有讨论过它们的基础或性质。具体来说，使用两种（或三种）态度价值观比许多真实语言示例所显示的要简单得多。在本文中，我们要问：情绪从何而来？我们关注人类态度的两个基本来源——持有者的非逻辑/情感偏好和持有者目标的实现。在探索了每个来源之后，我们提供了一个概念性的算法草图和示例，说明情感系统如何提供更丰富、更现实的文本情感描述。
The detection of expressions of sentiment in online text has become a popular Natural Language Processing application. The task is commonly defined as identifying the words or phrases in a given fragment of text in which the reader understands that the author expresses some person's positive, negative, or perhaps neutral attitude toward a topic. These four elements-expression words, attitude holder, topic, and attitude value-have evolved with hardly any discussion in the literature about their foundation or nature. Specifically, the use of two (or three) attitude values is far more simplistic than many examples of real language show. In this paper we ask: where do sentiments come from? We focus on two basic sources of human attitude-the holder's non-logical/emotional preferences and the fulfillment of the holder's goals. After exploring each source we provide a notional algorithm sketch and examples of how sentiment systems could provide richer and more realistic accounts of sentiment in text.