Prediction of human triglyceride concentration based on quantum genetic-partial least squares algorithm and...

Area of Science:

• Biomedical Engineering
• Analytical Chemistry
• Spectroscopy

Background:

• Accurate and convenient measurement of blood indicators is crucial for modern medical diagnostics.
• Triglyceride (TG) concentration is a key parameter in blood analysis with significant health implications.

Purpose of the Study:

• To develop a robust prediction model for triglyceride (TG) concentration in human serum.
• To investigate the utility of fluorescence spectra combined with advanced signal processing techniques for TG quantification.

Main Methods:

• Wavelet decomposition was employed to identify and reduce spectral noise, selecting 280 nm as the optimal excitation wavelength.
• Derivative spectroscopy and wavelet decomposition were used to minimize spectral multi-correlation.
• Prediction models were established using Quantum Genetic Algorithm and Partial Least Squares (PLS).

Main Results:

• Wavelet decomposition and derivative spectral analysis significantly reduced spectral multi-correlation and improved resolution.
• The developed prediction models achieved a Root Mean Square Error (RMSE) of 0.077 mmol/L.
• Information density analysis indicated that the 3rd layer detailed wavelet decomposition spectrum contains the most TG concentration information.

Conclusions:

• Fluorescence spectroscopy, enhanced by wavelet decomposition and derivative analysis, provides an effective method for predicting triglyceride concentration.
• The study demonstrates the potential for component concentration detection in complex multi-component systems.
• This approach offers a valuable reference for advancing component concentration detection in clinical diagnostics.