Rapid estimation of compost enzymatic activity by spectral analysis method combined with machine learning

Area of Science:

• Environmental Science
• Analytical Chemistry
• Biotechnology

Background:

• Compost enzymatic activity is a key indicator of compost quality and maturity.
• Traditional methods like fluorescein diacetate hydrolysis (FDA-HR) are time-consuming and labor-intensive.
• Developing rapid, non-destructive methods for assessing compost quality is crucial for sustainable waste management.

Purpose of the Study:

• To evaluate the feasibility of using visible near-infrared (VisNIR) diffuse reflectance spectroscopy (DRS) as a rapid and inexpensive alternative to traditional assays for predicting compost enzymatic activity.
• To compare the performance of different spectral preprocessing techniques and multivariate algorithms for predicting compost FDA-HR using VisNIR DRS data.

Main Methods:

• Compost samples from five facilities were analyzed using VisNIR DRS.
• Raw reflectance spectra underwent seven different preprocessing transformations.
• Six multivariate algorithms, including principal component analysis (PCA) and artificial neural networks (ANN), were employed to predict FDA-HR.
• ANN multilayer perceptron with Savitzky-Golay first derivative pretreatment showed the best performance.

Main Results:

• PCA effectively clustered samples by compost type but could not differentiate varying FDA levels.
• The ANN multilayer perceptron model achieved a residual prediction deviation (RPD) of 3.2, a validation R-squared (r²) of 0.91, and a root mean square error (RMSE) of 13.38 μg g⁻¹ h⁻¹.
• The ANN model successfully captured the complex, non-linear relationships between VisNIR spectra and compost enzymatic activity.

Conclusions:

• VisNIR DRS is a highly efficient and feasible technique for predicting compost enzymatic activity.
• This spectroscopic approach offers a significant advancement over traditional methods, enabling faster and more cost-effective compost quality assessment.
• The findings support the broader application of VisNIR DRS for monitoring both enzymatic and microbial activity in compost.