Prediction of total volatile basic nitrogen (TVB-N) in fish meal using a metal-oxide semiconductor electronic nose...

Area of Science:

• Food Science
• Analytical Chemistry
• Sensor Technology

Background:

• Total volatile basic nitrogen (TVB-N) is a key indicator of fish meal freshness, crucial for animal and human health.
• Accurate and rapid monitoring of TVB-N is essential for quality control in the fish meal industry.

Purpose of the Study:

• To develop a fast and accurate method for identifying TVB-N in fish meal using a self-developed electronic nose (e-nose).
• To establish a mapping relationship between e-nose sensor responses and TVB-N values for reliable freshness detection.

Main Methods:

• Variational Mode Decomposition (VMD) was used to decompose TVB-N variation curves into multiple frequency subsequences.
• Long Short-Term Memory (LSTM) models were applied to predict TVB-N values for each subsequence.
• The Sparrow Search Algorithm (SSA) optimized LSTM parameters (hidden units, learning rate, regularization coefficient) to enhance prediction accuracy.

Main Results:

• The VMD-LSTM model optimized by SSA demonstrated high accuracy in predicting TVB-N.
• The model achieved a coefficient of determination (R²) of 0.91.
• The root-mean-squared error (RMSE) and relative standard deviation (RSD) were 0.115 and 6.39%, respectively.

Conclusions:

• The developed method significantly improves the performance of e-nose for fish meal freshness detection.
• This approach offers a valuable reference for applying e-nose technology to quality detection in other materials.