Evaluating the plausible application of advanced machine learnings in exploring determinant factors of present pandemic: A case for continent specific COVID-19 analysis.
Evaluating the plausible application of advanced machine learnings in exploring determinant factors of present pandemic: A case for continent specific COVID-19 analysis.
复制标题
评估先进机器学习在探索当前流行病决定性因素中的合理应用：针对特定大陆的COVID-19分析案例。
DOI：
10.1016/j.scitotenv.2020.142723
复制
发表时间：
2021-04-15
期刊：
影响因子：
--
通讯作者：
Das DN
中科院分区：
文献类型：
--
作者：
Chakraborti S;Maiti A;Pramanik S;Sannigrahi S;Pilla F;Banerjee A;Das DN
冠状病毒病是一种新型的严重急性呼吸综合征(SARS，新冠肺炎)，由于其不可预测性和缺乏足够的药物，已成为全球关注的健康问题。机器学习(ML)模型可以有效地识别导致新冠肺炎总体死亡的最关键因素。ML模型在流行病学研究中的功能能力，特别是对新冠肺炎的研究，并没有得到实质性的探索。为了弥补这一差距，本研究采用了两个先进的最大似然模型，即。随机森林(RF)和梯度增强机(GBM)，用于执行回归建模并提供后续解释。经过五个连续的步骤进行分析：(1)确定相关的关键解释变量；(2)应用数据降维消除冗余信息；(3)利用ML模型衡量解释变量的相对影响力；(4)评价关键解释变量与新冠肺炎病例数和死亡数之间的相互关系；(5)时间序列分析，检验新冠肺炎的发病率和死亡率。在本研究考虑的解释变量中，空气污染、移民、经济和人口因素被发现是最显著的控制因素。由于探讨ML模型识别新冠肺炎关键决定因素的优越性的研究非常有限，本研究可以为未来的公共卫生研究提供参考。此外，本研究中使用的所有模型和数据都是开源的，可以免费获得，因此很容易实现重复性和科学性复制。
Coronavirus disease, a novel severe acute respiratory syndrome (SARS COVID-19), has become a global health concern due to its unpredictable nature and lack of adequate medicines. Machine Learning (ML) models could be effective in identifying the most critical factors which are responsible for the overall fatalities caused by COVID-19. The functional capabilities of ML models in epidemiological research, especially for COVID-19, are not substantially explored. To bridge this gap, this study has adopted two advanced ML models, viz. Random Forest (RF) and Gradient Boosted Machine (GBM), to perform the regression modelling and provide subsequent interpretation. Five successive steps were followed to carry out the analysis: (1) identification of relevant key explanatory variables; (2) application of data dimensionality reduction for eliminating redundant information; (3) utilizing ML models for measuring relative influence (RI) of the explanatory variables; (4) evaluating interconnections between and among the key explanatory variables and COVID-19 case and death counts; (5) time series analysis for examining the rate of incidences of COVID-19 cases and deaths. Among the explanatory variables considered in this study, air pollution, migration, economy, and demographic factor were found to be the most significant controlling factors. Since a very limited research is available to discuss the superiority of ML models for identifying the key determinants of COVID-19, this study could be a reference for future public health research. Additionally, all the models and data used in this study are open source and freely available, thereby, reproducibility and scientific replication will be achievable easily.
登录
查看更多内容
影响因子：
9.8
作者：
Adekunle, Ibrahim Ayoade;Onanuga, Abayomi Toyin;Ogunbanjo, Olakitan Wahab
通讯作者：
Ogunbanjo, Olakitan Wahab
影响因子：
3.3
作者：
Flaxman, Abraham D.;Vahdatpour, Alireza;Murray, Christopher J. L.
通讯作者：
Murray, Christopher J. L.
影响因子：
6.9
作者：
Chakraborti, Suman;Das, Dipendra Nath;Feng, Yongjiu
通讯作者：
Feng, Yongjiu
影响因子：
3.2
作者：
Harrison, RM;Jones, AM;Beswick, A
通讯作者：
Beswick, A
影响因子：
4.9
作者：
Azevedo, Leonardo;Pereira, Maria Joao;Soares, Amilcar
通讯作者：
Soares, Amilcar