Improved boosting and self-attention RBF networks for COD prediction based on UV-vis

Area of Science:

• Environmental Science
• Analytical Chemistry
• Spectroscopy

Background:

• Accurate water quality assessment is vital.
• Chemical Oxygen Demand (COD) is a key indicator.
• Traditional COD detection methods have limitations.

Purpose of the Study:

• To develop an optimized UV-vis spectroscopy model for COD detection.
• To address the challenge of turbidity interference in water samples.
• To enhance the accuracy and efficiency of COD measurement.

Main Methods:

• Utilized UV-vis spectroscopy to obtain absorption spectra from reservoir water samples.
• Developed an optimized boosting model with a turbidity compensation strategy.
• Implemented a self-attention mechanism within a radial basis function (RBF) network (saRBF framework).

Main Results:

• The proposed saRBF model achieved a coefficient of determination (R²) of 0.9267.
• The model demonstrated a root mean square error (RMSE) of 1.2669 and a mean absolute error (MAE) of 1.0097.
• The developed model outperformed existing COD measurement techniques.

Conclusions:

• The optimized boosting model effectively compensates for turbidity in UV-vis spectra.
• The saRBF framework offers a robust approach for accurate COD detection.
• This research provides a novel method for turbidity compensation and COD analysis in water quality monitoring.