Diabetes Classification using a MLP and an SVM

• Implemented several multi-layer perceptron models and an SVM for diabetes classification

• Used skorch along with grid search to tune MLP hyperparameters and scikit learn based grid search for SVM tuning

• Best performing MLP obtained validation accuracy of 0.84 and 0.87 accuracy on test set. Best performing SVM achieved accuracy of 0.77 on validation and 0.75 on test.

Packages

• Python Version: 3.8.11

• Libraries and Packages: numpy , pandas, seaborn, torch , sklearn , skorch

Specifications

Code, data and files for the above can be found in the following files:

• MLP: code for MLP model and corresponding visulisations
• SVM: code for SVM model and corresponding visulisations
• EDA: exploratory data analysis on the diabete dataset
• Data: PIMA Indians Diabetes dataset
• Best_models: comparing model performance between MLP and SVM