%0 Conference Paper %1 Nguyen:2011:AAP:2107636.2107651 %A Nguyen, Dong %A Smith, Noah A. %A Rosé, Carolyn P. %B Proceedings of the 5th ACL-HLT Workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities %C Stroudsburg, PA, USA %D 2011 %I Association for Computational Linguistics %K age author feature genre joint kallimachos learning model prediction text %P 115--123 %T Author Age Prediction from Text Using Linear Regression %U http://dl.acm.org/citation.cfm?id=2107636.2107651 %X While the study of the connection between discourse patterns and personal identification is decades old, the study of these patterns using language technologies is relatively recent. In that more recent tradition we frame author age prediction from text as a regression problem. We explore the same task using three very different genres of data simultaneously: blogs, telephone conversations, and online forum posts. We employ a technique from domain adaptation that allows us to train a joint model involving all three corpora together as well as separately and analyze differences in predictive features across joint and corpus-specific aspects of the model. Effective features include both stylistic ones (such as POS patterns) as well as content oriented ones. Using a linear regression model based on shallow text features, we obtain correlations up to 0.74 and mean absolute errors between 4.1 and 6.8 years. %@ 9781937284046

@inproceedings{Nguyen:2011:AAP:2107636.2107651, abstract = {While the study of the connection between discourse patterns and personal identification is decades old, the study of these patterns using language technologies is relatively recent. In that more recent tradition we frame author age prediction from text as a regression problem. We explore the same task using three very different genres of data simultaneously: blogs, telephone conversations, and online forum posts. We employ a technique from domain adaptation that allows us to train a joint model involving all three corpora together as well as separately and analyze differences in predictive features across joint and corpus-specific aspects of the model. Effective features include both stylistic ones (such as POS patterns) as well as content oriented ones. Using a linear regression model based on shallow text features, we obtain correlations up to 0.74 and mean absolute errors between 4.1 and 6.8 years.}, acmid = {2107651}, added-at = {2017-02-04T16:18:01.000+0100}, address = {Stroudsburg, PA, USA}, author = {Nguyen, Dong and Smith, Noah A. and Ros{\'e}, Carolyn P.}, biburl = {https://www.bibsonomy.org/bibtex/27df83622520e1e56d89f357280beb4f0/hotho}, booktitle = {Proceedings of the 5th ACL-HLT Workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities}, description = {Author age prediction from text using linear regression}, interhash = {bfbdb1b70bed31e3b2ac22ebfc5af860}, intrahash = {7df83622520e1e56d89f357280beb4f0}, isbn = {9781937284046}, keywords = {age author feature genre joint kallimachos learning model prediction text}, location = {Portland, Oregon}, numpages = {9}, pages = {115--123}, publisher = {Association for Computational Linguistics}, series = {LaTeCH '11}, timestamp = {2017-02-04T16:18:01.000+0100}, title = {Author Age Prediction from Text Using Linear Regression}, url = {http://dl.acm.org/citation.cfm?id=2107636.2107651}, year = 2011 }