EmoChannelAttn: Exploring Emotional Construction Towards Multi-Class Emotion Classification

Zongxi LI, Xinhong CHEN, Haoran XIE, Qing LI, Xiaohui TAO

Research output: Other Conference ContributionsConference Paper (other)Other Conference Paperpeer-review

Abstract

The current multi-class emotion classification studies mainly focus on enhancing word-level and sentence-level semantical and sentimental features by exploiting hand-crafted lexicon dictionaries. In comparison, very limited studies attempt to achieve emotion classification task from the emotion-level perspectives, which are to understand how the emotion of a sentence is constructed. Another limitation of existing works is that people assumed that emotion labels are relatively independent, neglecting the possible relations among different types of emotions.

In this paper, the team aims at exploring various fine-grained emotions based on domain knowledge to understand the construction details of emotions and the interconnection among emotions. To address the first issue, the team proposes a novel method named EmoChannel to capture the intensity variation of a particular emotion in time series by incorporating domain knowledge and dimensional sentiment lexicons. The resulting information of 151 available fine-grained emotions is utilised to comprise the sentence-level emotion construction. As for the second issue, the EmoChannelAttn Network is introduced to identify the dependency relationship within all emotions via attention mechanism to enhance emotion classification performance. The experiments demonstrate that the proposed method gains significant improvements compared with baseline models on several multi-class datasets.
Original languageEnglish
Publication statusPublished - 16 Dec 2020
EventThe 2020 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology - Virtual Conference
Duration: 14 Dec 202017 Dec 2020
http://wi2020.vcrab.com.au/

Conference

ConferenceThe 2020 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology
Abbreviated titleWI-IAT'20
Period14/12/2017/12/20
Internet address

Fingerprint Dive into the research topics of 'EmoChannelAttn: Exploring Emotional Construction Towards Multi-Class Emotion Classification'. Together they form a unique fingerprint.

Cite this