TY - GEN
T1 - Supervised and unsupervised feature selection for inferring social nature of telephone conversations from their content
AU - Stark, Anthony
AU - Shafran, Izhak
AU - Kaye, Jeffrey
N1 - Funding Information:
Dr. Clemens was an investigator for the ATHENA Trial (Mentor Corporation), is an investigator for the Motiva FDA Approval Trial (Establishment Labs), and was a former Allergan Consultant (2012–2015). None of the other authors have any financial relationships or affiliations to disclose. This study and the PROFILE Patient Registry is a project coordinated and supported by The Plastic Surgery Foundation.
PY - 2011
Y1 - 2011
N2 - The ability to reliably infer the nature of telephone conversations opens up a variety of applications, ranging from designing context-sensitive user interfaces on smartphones, to providing new tools for social psychologists and social scientists to study and understand social life of different subpopulations within different contexts. Using a unique corpus of everyday telephone conversations collected from eight residences over the duration of a year, we investigate the utility of popular features, extracted solely from the content, in classifying business-oriented calls from others. Through feature selection experiments, we find that the discrimination can be performed robustly for a majority of the calls using a small set of features. Remarkably, features learned from unsupervised methods, specifically latent Dirichlet allocation, perform almost as well as with as those from supervised methods. The unsupervised clusters learned in this task shows promise of finer grain inference of social nature of telephone conversations.
AB - The ability to reliably infer the nature of telephone conversations opens up a variety of applications, ranging from designing context-sensitive user interfaces on smartphones, to providing new tools for social psychologists and social scientists to study and understand social life of different subpopulations within different contexts. Using a unique corpus of everyday telephone conversations collected from eight residences over the duration of a year, we investigate the utility of popular features, extracted solely from the content, in classifying business-oriented calls from others. Through feature selection experiments, we find that the discrimination can be performed robustly for a majority of the calls using a small set of features. Remarkably, features learned from unsupervised methods, specifically latent Dirichlet allocation, perform almost as well as with as those from supervised methods. The unsupervised clusters learned in this task shows promise of finer grain inference of social nature of telephone conversations.
UR - http://www.scopus.com/inward/record.url?scp=84858961942&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84858961942&partnerID=8YFLogxK
U2 - 10.1109/ASRU.2011.6163973
DO - 10.1109/ASRU.2011.6163973
M3 - Conference contribution
AN - SCOPUS:84858961942
SN - 9781467303675
T3 - 2011 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2011, Proceedings
SP - 449
EP - 454
BT - 2011 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2011, Proceedings
T2 - 2011 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2011
Y2 - 11 December 2011 through 15 December 2011
ER -