Stylometric Analysis of Bloggers' Age and Gender

By Sumit Goswami, Sudeshna Sarkar, Mayur Rustagi. 2009

168 citations recorded by the source corpus.

We report results of stylometric differences in blogging for gender and age group variation. The results are based on two mutually independent features. The first feature is the use of slang words which is a new concept proposed by us for Stylometric study of bloggers. Slang is a non-dictionary word that has evolved with time due to its frequent and popular usage. For the second feature, we have analysed the variation in average length of sentences across various age groups and gender. These two features are then augmented with previous study results reported in literature for stylometric analysis of age and gender. The combined feature list enhances the accuracy by a remarkable extent in predicting age and gender. These experiments were done on a 20,000 blog corpus. Experimental results show that these features work well in detection of bloggers’ demography. However, gender determination is more accurate than age group detection over a data spread across all ages but the accuracy of age prediction increases if we sample data with remarkable age difference.

Subfields: Artificial Intelligence, Sociology and Political Science

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 Sociology and Political Science

Citations in the corpus

"How Old Do You Think I Am?" A Study of Language and Age in Twitter (2013, background): “…automatic age prediction has mostly been approached as a two-class or three-class classiﬁcation problem based on age spans with for example boundaries at 30 or 40 years (e.g. (Rao et al. 2010; Garera and Yarowsky 2009; Goswami, Sarkar, and Rustagi 2009)), thus corresponding to an etic approach.”
Demographic Factors Improve Classification Performance (2015, methodology): “For more information on Trustpilot as a resource, see Hovy et al. (2015a). We split each review into sentences, tokenize, replace numbers with a 0, lowercase the data, and join frequent bigrams with an underscore to form a single token.”
Computational Sociolinguistics: A Survey (2016, methodology): “com data, has been used in various studies on gender and age [7, 84, 93, 185, 227].”
External validity of sentiment mining reports: Can current methods identify demographic biases, event biases, and manipulation of reviews? (2014, unknown): “The list of papers found is here: [12,14–17,26,28,31–35,43–60].”
Developing Age and Gender Predictive Lexica over Social Media (2014, background): “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).”
The Limitations of Stylometry for Detecting Machine-Generated Fake News (2019, methodology): “In a related line of study, stylometry was applied to, rather than detecting the speciﬁc person, identifying characteristics of the author, such as age and gender (Goswami, Sarkar, and Rustagi 2009), political views (Potthast et al. 2018), or IQ (Abramov and Yampolskiy 2019).”
Exploring Demographic Language Variations to Improve Multilingual Sentiment Analysis in Social Media (2013, background): “…found that the majority of emotion-bearing features, e.g., emoticons, repeated letters, exasperation, are used more by female than male users, which is consistent with results reported in other recent work (Garera and Yarowsky, 2009; Burger et al., 2011; Goswami et al., 2009; Argamon et al., 2007).”
Overview of the Author Profiling Task at PAN 2013 (2013, background): “[12] added some new features as slang words and the average length of sentences, improving accuracy to 80.”
Emphasizing personal information for Author Profiling: New approaches for term selection and weighting (2018, background): “For example, [2] predicted age and gender of blogs’ authors by means of slang words and the length of sentences, whereas [27] and [28] used style and structural features such as the frequency of capital letters, words length, and number of words with ﬂooded characters (e.g., Heeeellooo), number of…”
Age Prediction in Blogs: A Study of Style, Content, and Online Behavior in Pre- and Post-Social Media Generations (2011, methodology): “We find that both stylistic and content features have a significant impact on age prediction and show that, for unseen blogs, we are able to classify authors as born before or after 1979 with 80% accuracy and born before or after 1984 with 82% accuracy.”
Why Gender and Age Prediction from Tweets is Hard: Lessons from a Crowdsourcing Experiment (2014, background): “, 2013), blogs (Rosenthal and McKeown, 2011; Goswami et al., 2009), chats in social networks (Peersman et al.”
It’s going to be okay: Measuring Access to Support in Online Communities (2018, unknown): “…the language of support based on the posting individuals, we include features known to be associated with age such as elongation and capitalization (Goswami et al., 2009; Barbieri, 2008), grammatical differences in sentence construction and length (Hovy and Søgaard, 2015), and a lexicon for age of…”
Extracting Human Temporal Orientation from Facebook Language (2015, background): “Others, such as the bottom three examples, might reference multiple temporal classes or not include clear verb tense and thus rely on the raters’ judgements for what is most dominant.”
Multitask Learning for Mental Health Conditions with Limited Social Media Data (2017, background): “ural language processing (NLP) to predict author characteristics such as age (Goswami et al., 2009; Rosenthal and McKeown, 2011; Nguyen et al., 2014), gender (Sarawgi et al.”