mrcolo / cs145covid

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

cs145covid

How to run the models: Install requirements shared by all models:

pip install sklearn
pip install pandas
pip install matplotlib

*Note, we are using the PIP python library manager

Feedforward NN:

Requirements:
Python3 version 3.9.1 or later

pip install pytorch-lightning
python3 ./regression.py

Predictions will be written to Team7.csv.

Simple Autoregression:

Requirements:
Python3 version 3.9.1 or later
statsmodels version 0.12.1 or later

Run commands:

pip install statsmodels
./runSimple

*Note: statsmodels is a large library and could take some time to install

Predictions will be written to simpleAR_output.csv.

Vector Autoregression:

Requirements:
Python3 version 3.9.1 or later
statsmodels version 0.12.1 or later

Run commands:

pip install statsmodels 
./runVAR

Predictions will be written to VAR_output.csv.

LSTM:

Requirements:
Python3 version 3.9.1 or later
Pytorch version 1.8.1 or later
Jupyter-Notebook version 6.3.0 or later

Run commands:

pip install torch
jupyter-notebook lstm.ipynb

After opening the Jupyter Notebook just run all the cells.

Predictions will be written to submission.csv.

Mape Scores (Kaggle Submission)

Neural Network: 1.158
Autoregression: 1.240
Vector Autoregression: 1.445
LSTM: 5.980

About


Languages

Language:Jupyter Notebook 87.5%Language:Python 12.5%