Knowledge Discovery and Scientometric Analysis of Conversational Artificial Intelligence (CAI) for Mental Health

Artificial intelligence (AI), is being used in the prediction, detection, and treatment of mental health care. Mental Health (MH) chatbots can interact with users in a conversational manner, providing them with information, resources, and emotional support serving as a source of information and resources. The quality of studying scientific literature and the publication patterns of authors, institutions, and countries is realized in this upshot through Scientometric Analysis (SA). This work applies SA and Semantic Knowledge Discovery (SKD) techniques to explore the knowledge and research trends of Conversational Artificial Intelligence (CAI) in the field of MH. The aim of the research is to discover the knowledge of CAI for MH research, to mutually validate the discovered knowledge using semantic and syntactic knowledge, and to uncover and analyse the research trends in several perspectives of CAI for MH. Methodical syntactic-based SA and SKD using the state-of-the-art pre-trained model BERT is carried out. This work establishes fourteen SA parameters along with seven diagnosis and dialogue parameters, consequently materializing pre-trained language model parameters for CAI in MH, which is a crucial deficit in contemporary research, to the best of our knowledge. Furthermore, the research trends are analysed based on countries, institutions, and keywords. This work offers valuable insights into the development and advancement of CAI in the MH domain. However, it is important to acknowledge that the findings are subjected to the database used for collection, which is, primarily, the Web of Science. This work contributes to the understanding of CAI in MH research and offers directions for future studies.