Novel approach for domain recognition of professional speech conversation using gaussi an naïve bayes and KNN classifier

Speech is the basic need of human being for communication. It exchanges useful information between two persons. Apart from contents information, some additional information like feelings, emotions, sentiments are also reflected in the communication. This paper proposes a novel and systematic approach for recognition of conversation into multiple domain's categories using Gaussaian naïve bayes and KNN classifiers. The categories of domain are defined through various real life professional conversations such as: Engineering, Medical, Research, Constitution & Law, Marketing, Religious, Social, and Sports. A professional conversation always belongs in any of these domain categories. Recognizing domain of conversation is a challenging research theme and can be useful for various applications based on human computer interaction.

We surveyed prior work and challenges presented by various researchers in the field of affective computing. Our system is capable enough to recognize the domain of professional conversation. The experimental results of proposed method show that it works efficiently with acceptable accuracy.