Survey on Ontologies for Affective States and Their Influences

Tracking #: 1418-2630

Authors: 
Rana Abaalkhail
Benjamin Guthier
Rajwa Alharthi
Abdulmotaleb El Saddik

Responsible editor: 
Harith Alani

Submission type: 
Survey Article
Abstract: 
Human behavior is impacted by emotion, mood, personality, needs and subjective well-being. Emotion and mood are human affective states while personality, needs and subjective well-being are influences on those affective states. Ontologies are means of representing real-world knowledge, such as human affective states and their influences, in a format that a computer can process. They allow researchers to build systems that harness affective states. By unifying terms and meanings, ontologies enable these systems to communicate and share knowledge with each other. In this paper, we survey existing ontologies on affective states and their influences. We also provide the psychological background of affective states, their influences and representational models. The paper discusses a total of 18 ontologies on emotion, one ontology on mood, one ontology on needs, and ten general purpose ontologies and lexicons. Based on the analysis of existing ontologies, we summarize and discuss the current state of the art in the field.
Full PDF Version: 
Tags: 
Reviewed

Decision/Status: 
Minor Revision

Solicited Reviews:
Click to Expand/Collapse
Review #1
Anonymous submitted on 13/Aug/2016
Suggestion:
Minor Revision
Review Comment:

This paper covers an important topic and provides a good introduction into the available ontologies for applications in affective computing. My opinion is that this review will be of substantial benefit to the broader semantic web community when considering affective computing, and does a decent job of discussing some of the main ontologies in this area.

The review is broad; however it could do with discussing some papers in more detail. For instance reference [77] Sykora et al. (2013)'s EMOTIVE ontology is mentioned in passing in the conclusion, where it really needs to be included in the section 4.2.1. Can authors please add this paper to table 1, and provide a brief paragraph of detail in section 4.2.1, as with the other ontologies. Also, the work on SenticNet by Cambria deserves some mention, could authors mention his work, at least in passing, e.g. "Cambria, Erik, Daniel Olsher, and Dheeraj Rajagopal. "SenticNet 3: a common and common-sense knowledge base for cognition-driven sentiment analysis." In Proceedings of the twenty-eighth AAAI conference on artificial intelligence, pp. 1515-1521. AAAI Press, 2014."

The review is well written with a good standard of English; however, I strongly recommend another proof-read by the authors. For instance on page 2 (second paragraph from the bottom of the page), it should say "a survey was carried out", and "However, this" rather than "However, This". Please check the rest of paper for similar mistakes. Also there are slight issues with formatting, table 3 and table 4 use a rather large font, where the font-size should be more in line with e.g. table 2.

Review #2
Anonymous submitted on 02/Oct/2016
Suggestion:
Accept
Review Comment:

On the whole this manuscript provides a useful and comprehensive background of the surveyed ontologies and formal systems. In general though, there is probably a little too much text devoted to explaining what ontologies are in the text.

The paragraph beginning with the sentence "Other researchers came up with other dimensions." in section 2.1 could use a couple of examples that illustrate the "other dimensions".

In Table 1, it is unclear what the difference between "discrete", "discrete model", and "discrete and dimensional" is . A paragraph explaining the Emotion Model descriptors used in the table would be very helpful.

Similarly Emotional Markup Language and EmotionML seem to be used interchangeably in the table and in various places in the text. Is EmotionML an abbreviation for the Emotion Markup Langugage. If so, then the name should be abbreviated the first time it appears in the text, e.g., Emotion Markup Language (EML), and the abbreviation should be used throughout afterwards.

Other than those relatively minor issues, this paper looks like a valuable addition to our knowledge base of available ontologies.

Review #3
Anonymous submitted on 18/Oct/2016
Suggestion:
Minor Revision
Review Comment:

The paper is a survey paper about ontologies for affective states. It provides an introduction of psychological research (and common models) for emotion, mood, sentiment, personality, needs and their relations to affective states. Afterwards, emotion related lexicons are described, before several ontologies are compared in the main section.

(1) Suitability as introductory text, targeted at researchers, PhD students, or practitioners, to get started on the covered topic.

+ introduction from psychology and semweb angles
+ structuring the area
+ comparison of several ontologies in tabular forms

(2) How comprehensive and how balanced is the presentation and coverage.

+ various covered and structured
(-) emotions in animals? (happy dog)
- relation to visual emotion/sentiment/aesthetics missing (beautiful sunset):
- e.g., http://www.ee.columbia.edu/ln/dvmm/vso/download/visual_sentiment_ontolog...
- Section 4.1/5: mention which personality trait classes you're referring to
- Section 4.2.1 2nd par: please rephrase, which ontology, which system
- Parts of Section 4 read like unconnected facts, try binding this more and relate to each other.

(3) Readability and clarity of the presentation.

- presentation: graphics look unprofessional and waste a lot of space
- Fig. 1: align text with arrows, make tips visible, use common RDF notation?
- Fig. 2 + 3: waste a lot of space
- Fig. 4: opening tag for emotion missing
- Fig. 6: subject missing, make this valid n3
- presentation: tables inconsistent (size, columns)
- Tables: 1st col should mention Ontology Name / prefix
- Table 1: typos (Llexicons/Llanguage), lines / whitespace would help for right col
- Table 2-5: missing col(s)
- Table 3, 4: size different
- presentation: large amount of typos, singular/plural errors, missing articles, please proofread! Some examples:
- "However,This paper"
- "Fiend of a Friend"
- "As a result, some rows are empty in Table 1." ("cells", no row nor column is empty)
- "WorNet-Affect"
- "The ontology conations"
- "develped"
- "machine(automatic"
- "which are represent in the"
- "an event take place"
(-) remove spaces before footnotes and ")"

(4) Importance of the covered material to the broader Semantic Web community.

+ Important to represent emotion / affective states in RDF

Other remarks:
- Introduction: missing ref to Gruber for "an explicit specification of a conceptualization"
- [77, 16, 75] should be introduced before the conclusion


Comments