Diabetes Detection: Predicting Type II Diabetes with Machine Learning Algorithm

This project aims to predict type II diabetes using machine learning models and identify the one with the best performance, followed by recommendations for prevention strategies. To enhance data quality and address abnormal values, three separate imputation methods were evaluated, as well as the impact of feature standardization on various models. Logistic regression with Ridge regularization, support vector machine, Random Forest and a simple neural network with one hidden layer were implemented with four models for comparison. Some important aspects of working with the dataset included feature engineering, and hyperparameter-tuning with cross-validation to get the best model fit. Feature importance analysis also provided valuable insights into the relationships between health metrics and diabetes risk.