AI & ML Internship - Task 3

Linear Regression using Titanic Dataset

Objective

The objective of this task is to implement and understand Linear Regression using Scikit-learn. The model is trained to predict passenger fare based on selected features from the Titanic dataset.

Tools Used

• Python
• Pandas
• NumPy
• Matplotlib
• Scikit-learn
• KaggleHub

Dataset

Titanic Dataset

Steps Performed

1. Data Loading

• Downloaded and loaded the Titanic dataset.
• Imported the dataset into a Pandas DataFrame.

2. Data Preprocessing

• Checked for missing values.
• Replaced missing values in the Age column using the median.
• Selected relevant features for regression.

3. Feature Selection

Input Features:

• Age
• Pclass

Target Variable:

• Fare

4. Train-Test Split

• Split the dataset into training and testing sets using an 80:20 ratio.

5. Model Training

• Created a Linear Regression model using Scikit-learn.
• Trained the model on the training dataset.

6. Prediction

• Generated fare predictions on the test dataset.

7. Model Evaluation

Evaluated model performance using:

• Mean Absolute Error (MAE)
• Mean Squared Error (MSE)
• R² Score

8. Visualization

Created the following plots:

• Actual vs Predicted Fare
• Residual Plot

9. Coefficient Interpretation

• Examined model intercept and feature coefficients.
• Analyzed the impact of Age and Passenger Class on Fare prediction.

Results

The Linear Regression model was successfully trained and evaluated. The evaluation metrics provided insight into prediction accuracy, while the visualizations helped assess model performance and residual distribution.

Key Concepts Learned

• Linear Regression
• Feature Selection
• Train-Test Splitting
• Model Training
• Prediction
• MAE, MSE, and R² Metrics
• Model Interpretation
• Residual Analysis

Outcome

Successfully implemented a Linear Regression model and evaluated its performance using standard regression metrics and visualizations.

Author

Submitted as part of the AI & ML Internship Program.