13th Annual Conference of the International Speech Communication Association

Portland, OR, USA
September 9-13, 2012

Personality Traits Detection Using a Parallelized Modified SFFS Algorithm

Clément Chastagnol (1,2), Laurence Devillers (1,3)

(1) Department of Human-Machine Interaction, LIMSI-CNRS, France
(2) University of Orsay PXI, France
(3) University of Sorbonne PIV, GEMASS-CNRS, France

We present in this paper a contribution to the INTERSPEECH 2012 Speaker Trait Challenge. We participated in the Personality Sub-Challenge, where the main characteristics of speakers according to the five OCEAN dimensions had to be determined based on short audio recordings solely. We considered the task as a general optimization problem and applied a parallelized version of a modified SFFS algorithm, wrapped around a SVM classifier, along with parameter tuning. Our system has yielded higher than baseline scores on all five dimensions, adding more than 20 percentage points to the recognition score of the Openness dimension.

Index Terms: personality detection, feature selection

Full Paper

Bibliographic reference.  Chastagnol, Clément / Devillers, Laurence (2012): "Personality traits detection using a parallelized modified SFFS algorithm", In INTERSPEECH-2012, 266-269.