|
Abstract:
|
In a leading service economy like India, services lie at
the very center of economic activity. Competitive organizations
now look not only at the skills and knowledge, but also at the
behavior required by an employee to be successful on the job.
Emotionally competent employees can effectively deal with
occupational stress and maintain psychological well-being. This
study explores the scope of the first two formants and jitter to
assess seven common emotional states present in the natural
speech in English. The k-means method was used to classify
emotional speech as neutral, happy, surprised, angry, disgusted
and sad. The accuracy of classification obtained using raw jitter
was more than 65 percent for happy and sad but less accurate for
the others. The overall classification accuracy was 72% in the
case of preprocessed jitter. The experimental study was done on
1664 English utterances of 6 females. This is a simple, interesting
and more proactive method for employees from varied
backgrounds to become aware of their own communication styles
as well as that of their colleagues' and customers and is therefore
socially beneficial. It is a cheap method also as it requires only a
computer. Since knowledge of sophisticated software or signal
processing is not necessary, it is easy to analyze |