Diabetes Prediction

Overview

A supervised machine learning project focused on binary classification to predict diabetes risk. Utilizes CDC data from Kaggle with over 250,000 patients and 22 features, including health and lifestyle factors.

Data Preprocessing

• Checked for missing values; dataset had none.
• Removed duplicate rows.
• Normalized the data and performed an 80:20 train-test split.

Feature Selection

• Conducted a chi-squared test to select features based on their scores.

Handling Imbalanced Data

• Applied undersampling and SMOTE to address data imbalance.

Model Building and Evaluation

• Experimented with various ML algorithms including Logistic Regression, KNN, SVM, and Random Forest.
• Random Forest achieved the best performance, with a recall of 0.89.