Analysis and areas modelling of Nasalised speech by a multivariable identification technique