This project aims to create a model to predict whether a patient has diabetes from analysing the patient's features. The dataset is processed in two different machine learning algorithms; decision tree and neural network. Hyperparameters are tuned for both of models to effectively train the model.
The performance of both models will be compared to derive to the best model. Furthermore, explainable AI using LIME library is used to understand how the neural network model make prediction. This project used some machine learning libraries such as TensorFlow Keras and SciKit Learn to make the classification.
- Tensorflow Keras
- Sklearn Decision Tree
- LIME Tabular Explainer
- Pandas
- Matplotlib
- Numpy
Jupyter notebook is required to run the code.
If the Jupyter notebook is in the local computer, copy this folder to your jupyter notebook path. Then, open the diabetes_classification.ipynb file in your local jupyter notebook. Run the code in the jupyter notebook.