Aplikasi yang dapat menentukan apakah seseorang boleh mendapatkan vaksin COVID-19 atau belum berdasarkan data riwayat penyakit.

Mata Kuliah Manajemen Proyek TI - 2021 Oleh:

• Stefannov
• M. Adisatriyo Pratama
• Surya Asmoro

• Aplikasi yang dapat menentukan apakah seseorang boleh mendapatkan vaksin COVID-19 atau belum berdasarkan data riwayat penyakit.
• Aplikasi berbentuk web yang dapat diakses melalui internet.
• Aplikasi ini menerapkan metode Machine Learning

• Muhammad Adisatriyo P :
  • Project Management
  • Build ML model
• Surya Asmoro:
  • Gather Data
  • Cleaning Data
  • Software Testing
• Stefannov:
  • Gather Data
  • Cleaning Data
  • Software Testing

Our goal for creating this project is to help the process of COVID-19 vaccine distribution faster so that we can reach the minimum 70% of total population to achieve national herd immunity for COVID-19.

• Python3
• Jupyter Notebook and Google Colab
• Scikit-Learn
• Google Docs
• Github
• Streamlit Framework
• Google Cloud Platform

Our problem is really simple binary Classification with just using decision tree classifier as our main algorithm and several parameters that is provided from scikit-learn can easily reach up to ~99% test accuracy.

from sklearn.model_selection import train_test_split

# split data
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, stratify=y, random_state=42)

# import library
from sklearn.tree import DecisionTreeClassifier

# model initialize with max_depth = 2
dt = DecisionTreeClassifier(max_depth=2, random_state=42)

from sklearn.metrics import accuracy_score

# fit training set
dt.fit(X_train, y_train)

# predict test set
y_pred = dt.predict(X_test)

# accuracy score
accuracy_score(y_test, y_pred)

0.9180327868852459

# Import library
from sklearn.model_selection import GridSearchCV

## hyperparameter
params_dt = {'max_depth':[1,2,3,4,6,8,10], 'min_samples_leaf':[0.0001, 0.001, 0.05, 0.1, 0.2], 'criterion':['gini', 'entropy']}

# performs GridSearchCV
grid_dt = GridSearchCV(estimator=dt, param_grid=params_dt, scoring='roc_auc', cv=5, n_jobs=1)

# fit data
grid_dt.fit(X_train, y_train)

print(f'Best Parameters : {grid_dt.best_params_}')
print(f'Best Score : {grid_dt.best_score_}')
print(f'Best Estimator : {grid_dt.best_estimator_}')

Best Parameters : {'criterion': 'gini', 'max_depth': 8, 'min_samples_leaf': 0.0001}
Best Score : 0.9989732930209121
Best Estimator : DecisionTreeClassifier(max_depth=8, min_samples_leaf=0.0001, random_state=42)

import pickle
pickle.dump(grid_dt.best_estimator_, open('clf.pkl', 'wb'))

We are using Google Cloud Platform to quickly deploy our web app. We are using Google App Engine and fetch the data from GitHub repository to build the web app.

Here is our web app URL: http://www.screening-apps.info/

• Python Docs: https://docs.python.org/3/
• Google Cloud Platform: https://cloud.google.com/
• Streamlit: https://streamlit.io/
• Scikit-learn: https://scikit-learn.org/