NLSIG-COVID19Lab

A playground for descriptive modelling and monitoring of the time-series COVID-19 pandemic growth with the nlogistic-sigmoid function.

nlogistic-sigmoid function (NLSIG) is a modern logistic-sigmoid function definition for modelling growth (or decay) processes. It features two logistic metrics (YIR and XIR) for monitoring growth from a two-dimensional (x-y axis) perspective.

Papers (Links)

• 

• NLSIG Conference Presentation Slides Best Student Paper at the 2nd African Symposium on Big Data, Analytics and Machine Intelligence and 6th TYAN International Thematic Workshop December 3-4, 2020.

• NLSIG Preprint

Data Source

World Health Organization

Getting Started : MATLAB

Matlab Toolbox Requirements - Optimization Toolbox - Statistics and Machine Learning Toolbox

MATLAB Release Compatibility - Compatible with R2020a and later releases

External Dependencies - None

MATLAB App: Installation

Provided is a MATLAB App to allow for easy use.

1. In the applet folder, double-click or right-click on the App Installer: NLSIG-COVID19Lab.mlappinstall

2. A confirmation dialog Install into My Apps pops up. Select Install

3. The App is then Installed, and can be accessed from the MY APPS section in the APPS tab on MATLAB's top panel or toolstrip. Hover on the NLSIG-COVID19Lab App icon to see the App details and install location.

The default install location is in the Add-Ons location and is specific to the platform. This can be viewed from the Home tab's Environment section, click Preferences > MATLAB > Add-Ons.

The install location then can be somewhere like this:

Windows - C:\Users\username\Documents\MATLAB\Add-Ons\Apps\NLSIGCOVID19Lab

Linux - ~/MATLAB/Add-Ons/Apps/NLSIGCOVID19Lab.

Mac - ~/Library/Application Support/MathWorks/MATLAB/Add-Ons/Apps/NLSIGCOVID19Lab

MATLAB App: Using

Click the NLSIG-COVID19Lab App icon to start the App.

See the screenshots below

To Start Modelling: Click

To Update the Local Database: Click

To View Available Country-Codes: Switch to the List:Country-codes Tab

To Set options for Modelling: Click on the Options Tab

Metrics: Interpretation

As at the 21st of February 2021:

For infections:

YIR = 0.822 [0.82, 0.843] indicates that the numbers are past the peak;

XIR = 2.95 [2.88, 2.96] indicates that this time is clearly a post-peak period.

For deaths:

YIR = 0.664 [0.61, 0.718] indicates that the numbers are no longer increasing and are recently past the peak for this phase;

XIR = 1.66 [1.41, 1.97] indicates that this time is an early post-peak period.

Metrics

R2GoF R2 Goodness of Fit

KSGoF Kolmogorov-Smirnov Goodness of Fit

KSdist Kolmogorov-Smirnov Distance Statistic

YIR Y-to-Inflection Ratio (Here Y = Infections or Deaths)

YIR < 0.5 indicates generally increasing motion of growth

YIR ~= 0.5 indicates generally that the increase has peaked.

YIR > 0.5 indicates generally reducing motion of growth

YIR ~= 0 indicates either that the growth is flattening or could be increasing.

XIR X-to-Inflection Ratio (Here X = Time in Days)

XIR < 1 indicates a pre-peak period

XIR ~= 1 indicates a peak-period.

XIR > 1 indicates a post-peak period.

XIR ~= 0 indicates either a post-peak period or an early pre-peak.

Frontend API

Examples of Frontend APIs available for this software package can be found in the:

1. examples_m_api folder and

2. examples_mlx_api folder

Saved Results

Saved model fit results and logistic metrics for infections and deaths can be found in the assets folder and measures folder

assets folder

Stores all graphics for the model fit of infections and deaths in a folder named by the last date time-stamp in the data. Graphics are individually saved using the country code.

For example: WDi.pdf and WDd.pdf respectively indicates the saved graphics of the COVID-19 infections and deaths model fit for the World to the last date time-stamp in the data in pdf format.

measures folder

Stores all estimated logistic metrics for infections and deaths till the last date time-stamp in the data in the infs and dths subfolders respectively.

Automated Tests

Automated Tests for the app and package functionalties can be found in the tests folder.

<!-- #### Example --><!-- Running 'querysingle.m' with the searchcode as WD gave the following model fit for the ongoing COVID-19 pandemic with respect to the last updated date of the data. -->

Contributions, Issues or Support

If you are:

• interested in dedicating the time to port to other languages or contribute to the software

• want to report issues or problems with the software

• seek miscellanous support

Then, you may contact me, by creating a new issue.

License

This work is free software under the BSD 3-Clause "New" or "Revised" License

Citation Details

Somefun et al., (2021). NLSIG-COVID19Lab: A modern logistic-growth tool (nlogistic-sigmoid) for descriptively modelling the dynamics of the COVID-19 pandemic process. Journal of Open Source Software, 6(60), 3002, https://doi.org/10.21105/joss.03002