Physics-Informed & Scientific Metrics

INFO

This folder contains evaluation metrics used in physics-informed and scientific machine learning.
These metrics assess physical consistency, equation compliance, and perceptual similarity in simulations and field data.

Physical Consistency Metrics

• Relative Error
• Energy Conservation Error

Equation-Based Metrics

• Residual Loss
• Physics-Informed Loss

Perceptual & Structural Metrics

• Structural Similarity Index (SSIM)