Diabetes Free

A machine learning ensemble model that predicts whether a patient has diabetes or not. Google Collaboratory was used to easily organize the code and data visualization.

Libraries used:

Data preprocessing and visualization:

• pandas, numpy, matplotlib and seaborn

Model developement:

• sklearn and xgboost.

The ensemble is composed of 3 models:

• Deep Neural Network (MLPClassifier from sklearn.neural_network)
• Random Forests (RandomForestClassifier from sklearn.ensemble)
• XGBoost (XGBClassifier from xgboost)

Each model's hyperparameters are individually fine-tuned by performing GridSearchCV (from sklearn.model_selection) and then the ensemble is built with VotingClassifier from sklearn.ensemble.

There is a detailed description of the dataset and it's features in the Google Collaboratory notebook.

Acknowledgements:

• Smith, J.W., Everhart, J.E., Dickson, W.C., Knowler, W.C., & Johannes, R.S. (1988). Using the ADAP learning algorithm to forecast the onset of diabetes mellitus. In Proceedings of the Symposium on Computer Applications and Medical Care (pp. 261--265). IEEE Computer Society Press.

• Tensorflow 2.0 Practical: https://www.udemy.com/course/tensorflow-2-practical

• Kaggle: https://www.kaggle.com/kumargh/pimaindiansdiabetescsv