Automatic detection of vocalized hesitations in Russian speech


2018. №6, 104-118

Vasilisa O. Verkhodanova @,
Vladimir V. Shapranov,
Irina S. Kipyatkova,
Alexey A. Karpov

St. Petersburg Institute for Informatics and Automation, Russian Academy of Sciences, St. Petersburg, Russia; @ vass.verkhodanova@gmail.com

Abstract:

The article is focused on the automatic detection of the most frequent speech disfluencies in Russian speech — hesitations. Authors describe the acoustic features of Russian hesitations as well as analyze the different methods of hesitation detection. Results of acoustic analysis have shown that hesitations in Russian speech tend to be centralized, and dependent on the speech genre influence the context differently. Experiments on computerized detection of hesitations in Russian speech confirmed the efficiency and adequacy of the approaches based on acoustic information alone. Support vector machines method yielded the best results with the weighted harmonic mean of precision and recall reaching 56 %.

For citation:

Verkhodanova V. O., Shapranov V. V., Kipyatkova I. S., Karpov A. A. Automatic detection of vocalized hesitations in Russian speech. Voprosy Jazykoznanija. 2018. No. 6. Pp. 104–118. DOI: 10.31857/S0373658X0002022-3.

Acknowledgements:

The research is supported by RFBR (projects No. 15-06-04465 and 18-07-01407), the Council for grants of the President of the Russian Federation (projects No. MK-1000.2017.8 and MД-254.2017.8), and the budget theme No. 0073-2018-0002.