Waku Jules, Oshinubi Kayode, Adam Umar Muhammad, Demongeot Jacques
IRD UMI 209 UMMISCO and LIRIMA, University of Yaounde I, Yaounde P.O. Box 337, Cameroon.
AGEIS Laboratory, UGA, 38700 La Tronche, France.
Diseases. 2023 Oct 3;11(4):135. doi: 10.3390/diseases11040135.
The objective of this article is to develop a robust method for forecasting the transition from endemic to epidemic phases in contagious diseases using COVID-19 as a case study.
Seven indicators are proposed for detecting the endemic/epidemic transition: variation coefficient, entropy, dominant/subdominant spectral ratio, skewness, kurtosis, dispersion index and normality index. Then, principal component analysis (PCA) offers a score built from the seven proposed indicators as the first PCA component, and its forecasting performance is estimated from its ability to predict the entrance in the epidemic exponential growth phase.
This score is applied to the retro-prediction of endemic/epidemic transitions of COVID-19 outbreak in seven various countries for which the first PCA component has a good predicting power.
This research offers a valuable tool for early epidemic detection, aiding in effective public health responses.
本文的目的是开发一种稳健的方法，以新冠疫情为例，预测传染病从地方病阶段向流行阶段的转变。
提出了七个用于检测地方病/流行病转变的指标：变异系数、熵、主/次主导频谱比、偏度、峰度、离散指数和正态性指数。然后，主成分分析（PCA）根据这七个指标构建一个得分作为第一个主成分，并根据其预测进入流行指数增长阶段的能力来评估其预测性能。
该得分应用于七个不同国家新冠疫情爆发的地方病/流行病转变的回顾性预测，其中第一个主成分具有良好的预测能力。
本研究为早期疫情检测提供了一个有价值的工具，有助于有效的公共卫生应对。
