BMI_Prediction

BMI values were predicted using the PIMA Indians dataset using linear regression and gradient descent. Dataset: https://www.kaggle.com/datasets/uciml/pima-indians-diabetes-database

The steps taken were as follows:

1. Load the data
2. Clean the dataset (seting target value, setting features, addressing missing values, and correcting types)
3. Normalize the data
4. Split the dataset into an 80% training and 20% testing set
5. Define an error calculation method
   
6. Define a gradient descent method
7. Define a MSE method (mean squared error)
8. Define a prediction method (dot product of weights and features)
9. Train the model using the methods defined above
10. Predict and tune the model

Cost vs Iteration Curve

Predicted vs Actual

PCA Dimentionality Reduction (8 features to 2)