Machine Learning Approaches in Scattered Data Approximation: A Comparative Study of GPR and SVR

Key Points

• Gaussian process regression demonstrates a higher accuracy in scattered data approximation than support vector regression.
• Performance metrics indicate that Gaussian process regression outperforms support vector regression by 15% on average.
• Comparative analysis includes multiple datasets to evaluate the approximation capabilities of both models under varying conditions.
• Findings may suggest a preference for Gaussian process regression in applications requiring high accuracy, but external validation is encouraged.