Long Short-Term Memory and its Applications: Study Cases on Diabetes, Malaria, Covid-19 and Multi-Disease

Abstract

The LSTM neural network was first proposed in 1997 to address the weakness of long-term RNN dependence. In appearance, the structure of the RNN network is such that it transmits even the information of the first words to the last words. But in practice, this does not happen and the RNN network has a weak long-term dependency. It is like saying that the RNN network does not have good long-term memory, in other words, it does not have long-term memory. The LSTM network has this long-term memory, meaning it can learn long-term dependency. LSTM network is used in various issues such as emotion analysis, language modeling, speech recognition, machine translation, text and image categorization, text production, natural language processing, time-series data, and intelligent health, etc. In this article, we review the various applications of the LSTM network in smart health with an emphasis on diabetes, malaria, Covid-19, and multi-disease.