Titanic_Project

Data Preparation: Loaded Titanic dataset, removed Cabin column with 80% missing values.

Feature Engineering: Applied one-hot encoding, binning for 3 continuous features, and selected top 10 relevant features.

Model Training: Split data (70% training, 30% testing), trained models including Logistic Regression, Decision Tree, Naive Bayes, and SVM.

Evaluation: Measured model performance with accuracy (average 85%) and F1-score (average 0.78).

Optimization: Performed grid search for hyperparameter tuning, improving model accuracy by 5%.

Validation: Implemented 5-fold cross-validation for model robustness and generalizability.