Neural network model based on an epidemic simulation, with the aim of finding the parameters to stem an epidemic https://youtu.be/sXSPE1favXY
This project utilizing a genetic algorithm and regression model to predict the number of infected individuals in a population. The algorithm takes into account various parameters such as the number of individuals who can cross, the percentage vaccinated, the percentage of individuals respecting quarantine, and the time of quarantine and travel. The results of the algorithm are then visualized and analyzed.
The whole project under unity
Doc -> https://docs.unity3d.com/Manual/index.html
All the models used in the project for each parameter are in .h5 format.
The set of databases for each parameter, all parameters are fixed except for the parameter treated by this database.
| Infected | Number | %Infected | %Quarantine | Infection Rate | Infection Time | Quarantine Time | Travel Time |
|---|---|---|---|---|---|---|---|
| 81 | 104 | 0.2 | 0 | 0.7 | 4 | 1 | 100 |
| ... | ... | ... | ... | ... | ... | ... | ... |
And all .pdf results from Google Colab for each parameters :
To reuse the project:
You can build the entire Unity project to get the simulation executable, and then use the code in GenerateData.py to create a database for each of the parameters. Finally, train the neural networks with Epidemic_Regression_Model.ipynb.
Or you can directly use the models here.
To finally use the genetic algorithm in Genetic_Algo.ipynb in order to obtain the results
Enjoy!