Investigating the Role of Emotion-Based Features in Author Gender Classification of Text

Montero, Calkin Suero; Munezero, Myriam; Kakkonen, Tuomo

doi:10.1007/978-3-642-54903-8_9

Cited by 16 publications

(11 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As mentioned in the introduction, social web communication styles vary on the social web in general (Burger, Henderson, Kim, & Zarrella, 2011;Mihalcea & Garimella, 2016;Volkova & Yoram, 2015), for expressing sentiment (Montero, Munezero, & Kakkonen, 2014;Thelwall, Wilkinson, & Uppal, 2010), and for evaluating products (Yang, Kotov, Mohan, & Lu, 2015). In many different (mainly offline) contests, males seem more inclined to discuss aspects of objects whereas females are more likely to refer to psychological and social issues (Newman, Groom, Handelman, & Pennebaker, 2008).…”

Section: Background: Algorithmic Biasmentioning

confidence: 99%

“…Analysts may implicitly assume that sentiment analysis results are unbiased because they are automatic but this is not necessarily true. Given the existence of clear gender differences in communication styles on the social web (Burger, Henderson, Kim, & Zarrella, 2011;Mihalcea & Garimella, 2016;Volkova & Yoram, 2015), including for expressing sentiment (Montero, Munezero, & Kakkonen, 2014;Thelwall, Wilkinson, & Uppal, 2010), interpreting sentiment (Guerini, Gatti, & Turchi, 2013) and discussing products (Yang, Kotov, Mohan, & Lu, 2015), gender biases in sentiment analysis seem likely. In other words, sentiment analysis algorithms may be more able to detect sentiment from one gender than from another so that, in a gender-mixed collection of texts, sentiment analysis results could over represent the opinions of one gender.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Gender bias in machine learning for sentiment analysis

Thelwall

2018

OIR

View full text Add to dashboard Cite

Purpose: This paper investigates whether machine learning induces gender biases in the sense of results that are more accurate for male authors than for female authors. It also investigates whether training separate male and female variants could improve the accuracy of machine learning for sentiment analysis. Design/methodology/approach: This article uses ratings-balanced sets of reviews of restaurants and hotels (3 sets) to train algorithms with and without gender selection. Findings: Accuracy is higher on female-authored reviews than on male-authored reviews for all data sets, so applications of sentiment analysis using mixed gender datasets will over represent the opinions of women. Training on same gender data improves performance less than having additional data from both genders. Practical implications:End users of sentiment analysis should be aware that its small gender biases can affect the conclusions drawn from it and apply correction factors when necessary. Users of systems that incorporate sentiment analysis should be aware that performance will vary by author gender. Developers do not need to create gender-specific algorithms unless they have more training data than their system can cope with. Originality/value: This is the first demonstration of gender bias in machine learning sentiment analysis.

show abstract

Section: Background: Algorithmic Biasmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Gender bias in machine learning for sentiment analysis

Thelwall

2018

OIR

View full text Add to dashboard Cite

show abstract

“…They present a word cloud-based technique to visualize results of DLA. Montero et al (2014) incorporated the feature attributes based on emotions, used the SVM classifier to test the results and obtained a recognition accuracy of 80%. Mukherjee and Liu.…”

Section: Related Workmentioning

confidence: 99%

The construction of Chinese microblog gender-specific thesauruses and user gender classification

Zhu

Cui

et al. 2018

Appl Netw Sci

View full text Add to dashboard Cite

Based on the statistical features, short text messages published by different gender users are different in terms of the words and semantics used. In this paper, two new features are constructed after constructing a gender-specific thesaurus. A new classification model is constructed by combining the traditional statistical features and the improved text implicitness feature. The experimental evaluation performed on the Sina Weibo dataset demonstrated the effectiveness of gender-specific thesaurus-based features, and the improved text implicitness feature improved the accuracy of gender classification to 84.7%.

show abstract

“…The WordNet-Affect, combines 1,316 words in 250 classes. For example, a positive feeling could fall into the class of 'liking', identified through words such as 'approval', 'sympathy' or 'friendliness' [41]. Additionally, text classification is supported by a number of disambiguation rules to exclude instances where words implying positive feelings are negated or an emotion bearing word fulfills a different role, such as the use of 'like' as a preposition, meaning 'similar to'.…”

Section: Concepts and Their Affective Implications: Emotional Profilimentioning

confidence: 99%

“…In general, the aim of characterizing the feelings present in a text can be achieved either through word-list associations (affective dictionaries and databases of common-sense knowledge) or machine learning [39]. For our purposes we used the SentiProfiler, an emotional analysis system described in [40,41]. The SentiProfiler uses an ontology, i.e.…”

Section: Concepts and Their Affective Implications: Emotional Profilimentioning

confidence: 99%

An Empirically Informed Taxonomy for the Maker Movement

2016

View full text Add to dashboard Cite

Abstract. The Maker Movement emerged from a renewed interest in the physical side of innovation following the dot-com bubble and the rise of the participatory Web 2.0 and the decreasing costs of many digital fabrication technologies. Classifying concepts, i.e. building taxonomies, is a fundamental practice when developing a topic of interest into a research field. Taking advantage of the growth of the Social Web and participation platforms, this paper suggests a multidisciplinary analysis of communications and online behaviors related to the Maker community in order to develop a taxonomy informed by current practices and ongoing discussions. We analyze a number of sources such as Twitter, Wikipedia and Google Trends, applying co-word analysis, trend visualizations and emotional analysis. Whereas co-words and trends extract structural characteristics of the movement, emotional analysis is non-topical, extracting emotional interpretations.

show abstract

Investigating the Role of Emotion-Based Features in Author Gender Classification of Text

Cited by 16 publications

References 33 publications

Gender bias in machine learning for sentiment analysis

Gender bias in machine learning for sentiment analysis

The construction of Chinese microblog gender-specific thesauruses and user gender classification

An Empirically Informed Taxonomy for the Maker Movement

Contact Info

Product

Resources

About