The Effectiveness of Disease Prediction in Enhancing Patient Satisfaction at the Community Level

Abstract

In contemporary society, trends in infectious disease occurrences, incidence, and prevalence are unknown in the majority of cases, complicating efforts geared towards disease prediction. In this study, the motivation was to contribute to infectious disease prediction through deep learning algorithms’ parameters’ optimization, with big data in the healthcare industry on consideration. Specifically, there was a comparison between the performances of the long-short term memory (LSTM) learning algorithm and deep neural network model with the performance of ARIMA (autoregressive integrated moving average) algorithm. Three infectious diseases were examined to discern model reliability and validity in under different experimental conditions. From the findings, this study established that the performance of LSTM and DNN frameworks is superior to that of ARIMA. Relative to chickenpox prediction, there was improvement by 19% and 25% after implementing LSTM and top-10 DNN models, respectively. Also, LSTM exhibited better accuracy while DNN exhibited more performance stability, especially with the spread of infectious diseases. The implication for the healthcare industry is that this study’s findings could be used to inform some of the ways in which reporting delays could be eliminated or minimized, improving on the current surveillance systems.