🛒 Car Seat Sales Prediction Using Regression Models

🔍 Overview

This project predicts car seat sales using machine learning and statistical modeling. By analyzing demographic and economic features, we aimed to identify key drivers of sales and build a model that can help optimize retail strategies.

🧭 Approach

We explored and prepared the dataset, applied regularization and ensemble methods, and validated the models using error metrics and cross-validation. LASSO and Random Forest were used to build interpretable and accurate predictive models.

⚙️ Methodologies

• LASSO Regression: For feature selection and regularization
• Random Forest: For capturing non-linear relationships and boosting accuracy
• Cross-Validation: To ensure model generalizability
• Evaluation Metric: Mean Squared Error (MSE)

🧰 Technologies

• Language: R
• Libraries: glmnet, MASS, dplyr, car, alr4
• Techniques: Regression Modeling, Feature Selection, Cross-Validation

💡 Key Learnings

• Learned how to build and compare multiple regression models
• Improved understanding of multicollinearity and variable selection
• Experienced real-world business forecasting using sales data

📈 Results

The LASSO model achieved a Mean Squared Error of ~21.93, identifying key variables such as room size, income level, and distance. The model balanced accuracy with interpretability, making it suitable for retail decision-making.