Predicting Covid-19 pandemic’s growth and decay trajectories

The current covid-19 pandemic has created deep crises affecting lives, livelihoods and economies at an unparallel and unprecedented scale in modern times. There is lot of fear and uncertainty due to spread of the virus. In such a scenario we need accurate predictions to plan effectively. At this stage pure statistics have failed us miserably, where in most efforts have been to fit the data to a curve at the same time the predictions have fallen short. The SIR (Susceptible, Infected, Recovered) epidemiological model has fallen flat, inaccurate in their predictions too.

These models were flawed because they did not take into account a wide spectrum of factors like genetic diversity, population immunity, mobility pattern, social contact dynamics, spatial structures, pathogen mutation and evolution, exposure time to pathogen, viral load, super spreaders, seasonality and endemicity. It is near impossible to collect enough and qualitative data regarding each of the fore-mentioned factors and develop a generic model, factoring in all uncertainties and make a forecast on epidemic spread.

It is concerning to see that these predictions have added to uncertainty and fear. The media has been religiously printing the obnoxious predictions.

In the month of May’20 our data science team put in effort to make a new model to predict the pandemic’s growth and decay trajectories.

We have developed a model that is able to predict the Pandemic growth and decay trajectory for 20 countries including India. The model was completed 4 weeks ago and is holding predictions with around 95% accuracy. Please find below the trajectory predictions for India.

GLC -- Gaussian and Cauchy Lorentz composite

The Red line – Total number of Indices a week

The Blue Line – The model prediction

Full text of non-peer reviewed research publication can be downloaded at: https://doi.org/10.1101/2020.06.07.20124966

For any queries please write to – [email protected]