欢迎来到OAJRC平台! 英文版

aam@oajrc.org
首页 > 精选文章 > 精选文章

Advances in international applied mathematics. 2021; 3: (1) ; 10.12208/j.aam.20210001.

An Approach to Calibrate Basic Reproduction Number Based on Machine Learning and SIR Model
一种基于机器学习和SIR模型的基本再生数标定方法

作者: Angela Yuan*

THE HOCKADAY SCHOOL

*通讯作者:Angela Yuan,单位:THE HOCKADAY SCHOOL;

引用本文: Angela Yuan. 一种基于机器学习和SIR模型的基本再生数标定方法[J]. 国际应用数学进展, 2021, 3(1) : 1-4.
Published: 2021-01-21

摘要

基本再生数是判断一种流行病发展趋势的首要指标,对于制定应对疾病的政策具有既根本又直观的参考意义。本文采用SIR模型对流行病进行建模,使用美国COVID-19的数据对SIR模型中的参数进行基于机器学习的反演,并使用反演获得的参数标定基本再生数。在基于机器学习的参数反演中,使用SIR模型解变量的实际数据构造了一组新的特征变量,代替模型的解变量本身用作神经网络的输入变量,获得了更为满意的反演结果。由此获得的基本再生数,可以用来对美国COVID-19的整体发展趋势进行基本合理的解释。

关键词: 机器学习,SIR模型,基本再生数

Abstract

The basic reproduction number provides an overall measure of the potential for transmission of an infection within a population. In this paper, the SIR model is adopted as to investigate the COVID-19 data to find the transmission coefficient and the average duration of infectiousness in the model, thus the basic reproduction number is calibrated. Instead of the original data, a new set of features/variables calcualted from these data are used as the input to train the neural network based on state of the art machine learning. Significant improvement in the accuracy is achieved using this new approach. The basic reproduction number obtained from US COVID-19 data is quite reasonable in explanations for the general trend of the COVID-19 pandemic in USA.

Key words: Machine Learning, SIR Model, Basic Reproduction Number

参考文献 References

[1]

[2]

[3]

[4]

[5]

[6]

  • 123

    总浏览量