The GMM classifier returns probabilities (the so-called scores) that the tested utterance belongs to the GMM model while the identification of emotion (or gender) i* is given by the maximum overall probability for the given emotion (gender) i * = arg max 1 ≤ i ≤ N score T, i, (3).
The second interaction term is parental educational level and gender (i.e., parental educational level × gender).
The first interaction term is family economic conditions and gender (i.e., family economic conditions × gender).
The results are independent of gender, i.e., women and men show a similar correlation in these traits.
The third interaction term is differences between urban and rural and gender (i.e., urban-rural differences × gender).
This group was also balanced by gender (i.e. five boys, five girls) and participants were again chosen at random.
However some causes are more frequent in the paediatric population (like appendicitis or adenomesenteritis) or are strictly related to the gender (i.e. gynaechologic causes).
The sample was stratified by age and gender, i.e., each age group included 2,000 persons of each gender, aged 30 60 years, while the remaining age groups included 1,000 persons of each gender.
