Intelligent HIV/AIDS FAQ Retrieval System using Neural Networks

Abstract

In this paper, we present an HIV/AIDS FAQ retrieval using Artificial Neural Network (ANN). One of the challenges in FAQ retrieval systems is the mapping of user queries to appropriate questions in the FAQ repository so that the corresponding answer can be retrieved as an answer to the user query. To address this issue, a number of approaches have been proposed, most of which are based on traditional information retrieval techniques. In this paper we discuss our approach based on Neural Network which maps user queries to existing questions in the FAQ repository using Multi-layered Feedforward Neural Network architecture (with back- propagation training). One of the advantages of the Neural Network approach is its ability to learn based on a given training data set and then map user queries to FAQ questions based on its training. For the implementation of our neural network based FAQ Retrieval, we have experimentally determined the appropriate neural network parameters using MATLAB. To evaluate the effectiveness of the Neural Network based FAQ retrieval, we compared the performance of the system with keyword-based FAQ retrieval system. The results show a better recall and rejection of the Neural Network based approach. The experimental results demonstrate that the Neural Network approach can effectively improve the performance of the HIV/AIDS FAQ retrieval system.