Market Segmentation and Classification of Vehicles Using Clustering and Machine Learning Models

Clustering Analysis and Model Performance Summary

Clustering Insights

The car dataset was analyzed using clustering techniques to categorize vehicles into distinct market segments. The identified clusters provide insights into different target markets and strategies:

1. Cluster 0 (Economical and Compact)

   • Target Market: Budget-conscious buyers or urban drivers.
   • Marketing Strategy: Emphasize affordability, fuel efficiency, and compact size. Highlight the value for money and low operating costs to appeal to budget-conscious consumers and first-time buyers.
   • Characteristics: High fuel efficiency, smaller engine sizes, and lower vehicle weights suggest compact or subcompact vehicles with affordable pricing.

2. Cluster 1 (High-Performance and Large)

   • Target Market: Affluent consumers seeking performance or luxury.
   • Marketing Strategy: Focus on performance, luxury features, and high-status appeal. Utilize premium marketing channels and emphasize exclusive features.
   • Characteristics: Larger engine sizes, higher horsepower, and greater vehicle dimensions indicate high-performance or luxury vehicles, such as SUVs or large sedans, with premium pricing.

3. Cluster 2 (Balanced Performance)

   • Target Market: Mid-range buyers looking for a balance of performance and comfort.
   • Marketing Strategy: Highlight a mix of performance, comfort, and practicality. Appeal to consumers who want a versatile vehicle offering good value.
   • Characteristics: Mid-range engine sizes and horsepower with moderate pricing suggest a balance between performance and fuel efficiency, suitable for mid-sized or larger compact cars.

4. Cluster 3 (Luxury and High-End)

   • Target Market: High-income consumers seeking top performance and luxury.
   • Marketing Strategy: Emphasize luxury, advanced technology, and high performance. Target high-end markets with exclusive promotions and premium experiences.
   • Characteristics: Very high horsepower and engine sizes with larger dimensions indicate luxury SUVs or high-end sports cars, with the highest MSRP and invoice values.

Model Performance Comparison

A comparison of various classification models was conducted to determine their effectiveness in predicting car classes:

• Logistic Regression and Random Forest both achieved perfect training accuracy and high test accuracy, indicating robust performance across training and testing datasets.
• Support Vector Machine (SVM) exhibited near-perfect training accuracy and the highest test accuracy, making it the top performer among the evaluated models.
• K-Nearest Neighbors (KNN) demonstrated good test accuracy with slightly lower training accuracy, suggesting some resistance to overfitting.
• Decision Tree achieved perfect training accuracy but showed significantly lower test accuracy, indicating a tendency to overfit the training data.

Hyperparameter Tuning Results

Hyperparameter tuning was performed to optimize model performance:

• Number of Fits: 4800 (5 folds for each of 960 candidate configurations)
• Best Parameters:
  • C: 1
  • coef0: 0
  • degree: 2
  • gamma: 'scale'
  • kernel: 'linear'
• Best Cross-Validation Accuracy: 0.9765
• Test Accuracy of the Best Model: 0.9884

The optimal model configuration, using a linear kernel with the specified hyperparameters, demonstrated high performance and accuracy on the test set, reflecting its strong generalization capabilities.

Summary

The clustering analysis provided valuable insights into market segments and targeted strategies for different types of vehicles. The evaluation of classification models revealed that the Support Vector Machine model achieved the highest accuracy, making it the most effective model for predicting car classes. The results from hyperparameter tuning further validated the model's performance, ensuring its robustness and reliability for deployment.