Exploring Demographic Language Variations to Improve Multilingual Sentiment Analysis in Social Media

By Svitlana Volkova, Theresa Wilson, David Yarowsky. 2013

180 citations recorded by the source corpus.

Different demographics, e.g., gender or age, can demonstrate substantial variation in their language use, particularly in informal contexts such as social media. In this paper we focus on learning gender differences in the use of subjective language in English, Spanish, and Russian Twitter data, and explore cross-cultural differences in emoticon and hashtag use for male and female users. We show that gender differences in subjective language can effectively be used to improve sentiment analysis, and in particular, polarity classification for Spanish and Russian. Our results show statistically significant relative F-measure improvement over the gender-independent baseline 1.5% and 1% for Russian, 2% and 0.5% for Spanish, and 2.5% and 5% for English for polarity and subjectivity classification.

Subfields: Artificial Intelligence, Human-Computer Interaction

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Human-Computer Interaction

Citations in the corpus

The Social Impact of Natural Language Processing (2016, background): “, 2015b), and the characteristics in turn can be detected by and influence the performance of our models (Mandel et al., 2012; Volkova et al., 2013; Hovy, 2015).”
Predictive Biases in Natural Language Processing Models: A Conceptual Framework and Overview (2019, background): “same vein, Webster et al. (2018) provide a genderbalanced training corpus for coreference resolution.”
Challenges of studying and processing dialects in social media (2015, background): “A number of recent papers have explored social media with respect to sociolinguistic and dialectological questions (Rao et al., 2010; Eisenstein, 2013; Volkova et al., 2013; Doyle, 2014; Hovy et al., 2015; Volkova et al., 2015; Johannsen et al., 2015; Hovy and Søgaard, 2015; Eisenstein, to appear).”
Demographic Factors Improve Classification Performance (2015, background): “Volkova et al. (2013) also explore the inﬂuence of gender and age on text-classiﬁcation.”
Tagging Performance Correlates with Author Age (2015, background): “In the light of recent research (Volkova et al., 2013; Hovy, 2015; Jørgensen et al., 2015), we explore the hypothesis that these biases transfer to NLP tools induced from these resources.”
Women’s Syntactic Resilience and Men’s Grammatical Luck: Gender-Bias in Part-of-Speech Tagging and Dependency Parsing (2019, background): “While data bias can be overcome by accounting for demographics, and can even improve classification performance (Volkova et al., 2013; Hovy, 2015; Bolukbasi et al., 2016; Benton et al., 2017; Zhao et al., 2017; Lynn et al., 2017), there is still little”
Developing Age and Gender Predictive Lexica over Social Media (2014, methodology): “Seeking out-of-domain data, we downloaded age and gender annotated blogs from 2004 (Schler et al., 2006) (also used in Goswami et al., 2009) and gender labeled tweets (Volkova et al., 2013).”
Studying User Income through Language, Behaviour and Affect in Social Media (2015, background): “Automatically inferred user characteristics can enable large-scale social science studies or assist applications such as targeted advertising, polling across different demographics or sentiment analysis [15].”
Predicting Twitter User Demographics from Names Alone (2018, methodology): “Twitter : The dataset created by Burger et al. (2011) (and processed and released by Volkova et al. (2013) 4 ) provides us with 58,046 users, 30,364 female and 27,682 male.”
Social and Emotional Correlates of Capitalization on Twitter (2018, methodology): “Unsurprisingly, genders differ in their use of emotive language as well (Volkova et al., 2013; Hovy, 2015).”
Square One Bias in NLP: Towards a Multi-Dimensional Exploration of the Research Manifold (2022, background): “%) is done only on English; however, ways of expressing sentiment (Volkova et al., 2013; Yang and Eisenstein, 2017; Vilares et al., 2018) and visually grounded reasoning (Liu et al., 2021a; Yin et al., 2021) do vary across languages and cultures.”
The Enemy in Your Own Camp: How Well Can We Detect Statistically-Generated Fake Reviews – An Adversarial Study (2016, background): “…of) start token(s), n is the size of the Markov horizon, and However, extra-linguistic information, if available, can improve classiﬁcation performance (Volkova et al., 2013; Hovy, 2015), and fake-review detection models often also exploit meta-information about the author and their behavior…”
Motivating Personality-aware Machine Translation (2015, background): “Such an approach has already shown to be useful for several NLP tasks (Volkova et al., 2013; Hovy, 2015).”
Equity Beyond Bias in Language Technologies for Education (2019, background): “…DMello and Graesser, 2012; Woolf et al., 2009), but these algorithms are known to produce discriminatory results, with disparate outcomes by gender (Volkova et al., 2013), race (Kiritchenko and Mohammad, 2018), and age (D´ıaz et al., 2018), perpetuating a quantiﬁable trend of disproportionate…”
Why Gender and Age Prediction from Tweets is Hard: Lessons from a Crowdsourcing Experiment (2014, background): “These differences have also been exploited to improve sentiment classification (Volkova et al., 2013) and cyberbullying detection (Dadvar et al.”
The Importance of Modeling Social Factors of Language: Theory and Practice (2021, background): “…various approaches have shown the positive impact of incorporating speaker characteristics into NLP applications, either as explicit features (Volkova et al., 2013), through conditional embeddings (Hovy, 2015; Lynn et al., 2017), or via neural models for multi-task learning (Ben-ton et al.,…”
Computational Sociolinguistics: A Survey (2016, methodology): “[259] studied gender differences in the use of subjective language on Twitter.”