Speech Acoustic Features Characterising Individuals with High and Low Public Trust

Trust has long been a focus of research in the areas of psychology and marketing. However, it has not been studied extensively in computer science and engineering. In this study, the notion of trust was investigated from the perspective of statistical analysis and automatic machine-based classification of speech. A gender-balanced database of speech samples representing individuals with low and high public trust was created and analyzed statistically to determine acoustic speech parameters correlated with trust. The effect of gender was analyzed. An automatic trust classification based on acoustic speech parameters was conducted using a neural network.