Long Short-Term Memory based RNN for COVID-19 disease prediction

被引：1

作者：

Bahri, Safa ^{[1
]}

Kdayem, Moetez ^{[2
]}

Zoghlami, Nesrine ^{[1
]}

机构：

[1] Natl Sch Engn Tunis, LTSIRS Lab, Tunis 1002, Tunisia

[2] Natl Sch Engn Sousse, Sousse, Tunisia

来源：

2021 22ND IEEE INTERNATIONAL CONFERENCE ON INDUSTRIAL TECHNOLOGY (ICIT) | 2021年

关键词：

Artificial Intelligence; LSTM; LightGBM; Forecasting models;

D O I：

10.1109/ICIT46573.2021.9453534

中图分类号：

T [工业技术];

学科分类号：

08 ;

摘要：

Currently, the global health system is suffering from an overwhelming issue affecting a large number of individuals all around the world. The novel coronavirus, called COVID-19, has continued to claim more than one million lives. In such cases, it is of vital importance to develop alternatives addressing this health issue and saving more lives. Artificial Intelligence were among the efficient tools that can address this global threat. In this study, we propose to test a recurrent neural network named Long Short-Term Memory (LSTM-RNN) for estimating the number of future fatality cases in USA, India and Italy. Our experimentations proved the effectiveness of LSTM-RNN in predicting the number of deceased cases with minimum of loss ranging from 1.37% to 2.7%.

引用

页码：901 / 906

页数：6

共 45 条

[11] Geographic Differences in COVID-19 Cases, Deaths, and Incidence - United States, February 12-April 7, 2020 [J].

MMWR-MORBIDITY AND MORTALITY WEEKLY REPORT, 2020, 69 (15) :465-471

[12] Estimation of COVID-19 prevalence in Italy, Spain, and France [J].