Skip to content

Latest commit

 

History

History
36 lines (28 loc) · 1.71 KB

README.md

File metadata and controls

36 lines (28 loc) · 1.71 KB
Raw

COVID-19

This repository is B.Tech. major project on COVID-19 Global and India Forescast.

The project covers a study of COVID-19's behaviour/trend in global regions, and indian regions along with forecasting deaths and positive cases with various methods.

The study and forecast on world level data is done for duration of Jan 22, 2020 to April 19, 2020, while on indian-region level data is done till May 19, 2020.

EDA

EDA for both of the can be found on the notebook in repo.

Forecasting

For the former following methods were used to forecast:

  • SARIMA
  • Facebook's Prophet Model
  • Linear Regression
  • LightGBM
  • Exponential Smoothing using Holt-Winters' Method or Triple Exponential smoothing

For Indian regions, following methods were used for forecasting:

  • Exponential Smoothing using Holt-Winters' Method or Triple Exponential smoothing
  • Facebook's Prophet Model
  • LightGBM
  • Linear Regression
  • Polynomial Regression
  • LASSO Regularisation
  • Ridge Regression

Link for the codes of forecasting (Notebook)

Results

The results of the study and forecasts can be seen in the below public tableau workbooks:

Presentation material

The repo contains the complete Powerpoint Presentation professionally built by me, and the tableau work books here