Categorization of Arabic Medical Questions Using a Deep Learning Approach

Recently, The use of AI in the medical domain has been growing rapidly, due to its capacity to shift several medical tasks, such as clinical decision-making, medical research, analysis, data processing, diagnosis, etc. In this regard, IA can improve the diagnosis using an automated categorization of patients’ questions. Moreover, it can ameliorate the quality of the provided services. So, medical question categorization is a key aspect of healthcare question-answering systems that use natural language processing (NLP). In this paper, an automated Arabic medical question categorization model is developed. This model is built based on the concatenation of two Deep Learning algorithms, namely, CNN and LSTM, and two methods of pre-trained word representations, general and specialized. Hence, to validate the model, a dataset of 22500 questions categorized into 9 specialties is used to train and verify its performance using the related performance metrics, such as accuracy, precision, recall, and F1-score. The obtained results of the proposed model demonstrate its potential to considerably improve the performance of Arabic medical questions categorization.