Forecast COVID-19 risk in any geographic area using a predictive neural network.
The code in this repository was a software engineering project in the GSMST engineering fair.
CRNN is a perceptron neural network designed for customizable COVID-19 forecasting.
Risk Table of All 50 States
Risk Table of Top 20 Most Populus Counties in Georgia
A dataset in a spreadsheet would look like this, having labels and text corresponding to numerical data.
| Geographical Area | Factor1 | Factor2 | Factor3 | Factor n... | Confirmed COVID-19 Deaths |
|---|---|---|---|---|---|
| area1 | # | # | # | # | x... |
| area2 | # | # | # | # | y... |
| area3 | # | # | # | # | z... |
Training and prediction datasets passed into the network must be stripped of all text as shown below.
A dataset may have any number of numerical input factors, but the last output column must always be the number of confirmed COVID-19 deaths in the respective geographic area.
This is becasue the training algorithm finds the impact correlation of each feature on the resulting deathcount of an area, being able to predict risk of new geographic areas.
| # | # | # | # | x... |
|---|---|---|---|---|
| # | # | # | # | y... |
| # | # | # | # | z... |
Datasets used for training, testing, and diagnosis throughout the course of this project can be found here.
- 1st Place in GSMST Science and Engineering Fair
- 1st Place in Gwinnett County Science and Engineering Fair