HEALTHCARE QUESTION AND ANSWER SYSTEM BASED ON SEQUENCE TO SEQUENCE MODEL

Abstract

The healthcare sector is one of the most important domains that impacts the entire global population and is closely linked to the development of any country. There are millions and billion pieces of healthcare information available, but making the right information accessible when needed is very important. The advent of Question Answering System has been applied as a promising solution and an efficient approach for retrieving significant healthcare information easily and time saving. The deep learning algorithms are used to train the data and bring output within a specific range by using statistical analysis. Recurrent Neural Network (RNN) based Sequence-to sequence (Seq2Seq) model is one of the most commonly researched model to implement artificial intelligence question answering system. This system has been implemented using neural network where it use bidirectional RNN as encoder and Luong Attention RNN as decoder. The system presented a healthcare question answering system to provide healthcare information based on three attention sequence to sequence models: general, dot and concat with three sub datasets: NCI, NIHSeniorHealth and NHLBI in MedQuAD.