This is a repository for implementing, describing, and testing a Single Layer Perceptron for predicting diabetes. The results are in form of a short conference paper. I achieved an accuracy of 80.58%. In the project i used 5 fold cross validation with shuffling for determining the ideal learning rate. I used the PIMA Indians dataset .The training was evaluated using evaluation metrics such as Accuracy, Execution and AUC on various learning rates. For a detailed analysis of the results, please refer to paper & code. Below are some of the results from the paper.
We use a very basic implementation of SLP(Single Layer Perceptron) in classifying if the patient is diabetic or not.
The project is implemented using a jupyter notebook, so its fairly straightforward to download it directly to retest. Notebook
The code utilizes the python packages as such
- matplotlib,
- seaborn,
- pandas,
- numpy,
- scikit-learn,
- Tensorflow.
Distributed under the MIT License. See LICENSE for more information.
Project Link:Project
Kaggle Notebook:
Collab Notebook:
