Interpretable framework based on stepwise virtual sample generation for accelerating accurate PCDD/Fs prediction and assessing emission control strategies under small-sample conditions

Key Points

• Prediction models enhanced accuracy for PCDD/Fs under small-sample conditions, improving control strategies.
• A key improvement was a 30% increase in prediction accuracy using stepwise virtual sample generation.
• The framework analyzes emission control strategies through advanced algorithms and predictive modeling.
• This method may enable more effective regulation of emissions in emerging areas, requiring further validation.