Malaysia Covid-19 Forecast using Generalised Logistic Function (Updated 14-Apr-2020)

In a previous article (Read more here), a Simple Logistic Function (SLF) was chosen to model the Covid-19 growth trend in Malaysia at the end of 1^st phase Movement Control Order (MCO).  During that time, the SLF model appeared to be adequate as the goodness-of-fit of the curve was reasonable.  However, at the end of the 2^nd phase of MCO (14-Apr-2020), the SLF model is no longer adequate to explain the development of Covid-19 in Malaysia.

A more robust model, Generalised Logistic Function (GLF) is now needed (Read more here), as suggested by some international research papers (Read more here).  There are few reasons why the SLF model is inadequate to predict the growth of the Covid-19.  Firstly, the Covid-19 development is not happened in a closed-system.  SLF is commonly used in studying the growth of bacteria in laboratory.  In the Covid-19 case, although MCO is implemented, it is not a true closed-system, there are leakages that will impact the growth pattern such as test capacity, asymptomatic patients, previous undisclosed linked clusters, and MCO violations.  Secondly, the previous model might appear good due to insufficient data points.  As time passed, more data are now available to show the actual trend of the development.

The GLF, in mathematical form, is

The constants A, K, C, Q, B and v are determined by minimizing the sum of square of the 21-Days rate of change between the actual cumulative cases and the GLF.  Graph 1 is the cumulative positive cases while Graph 2 is the 21-Days rate of change.  From the graphs, GLF (orange curve) fitted very well to actual data (blue curve).  Meanwhile, SLF (grey curve) is poorly fitted as at 14-Apr-2020.

Based on the fitted GLF, the predicted total cases are around 6200 at the middle of June 2020.  This number is derived purely using quantitative approach.  It does not factor in traffic movement, new vaccine development and other qualitative measures.  Nevertheless, the number could be further reduced if we abided to the MCO, and practice good social distancing.  Stay@Home!