Selective iteratively reweighted quantile regression for baseline correction.

Area of Science:

• Analytical Chemistry
• Chemometrics
• Signal Processing

Background:

• Baseline drift in analytical signals complicates data analysis, particularly in multivariate applications.
• Existing baseline correction methods often require manual intervention or lack robustness, especially for low signal-to-noise ratios.
• Accurate baseline correction is crucial for reliable qualitative and quantitative information extraction.

Purpose of the Study:

• To develop a novel, automated baseline correction algorithm.
• To address limitations of existing methods, such as user dependency and variability.
• To improve the accuracy and clarity of analytical signal analysis.

Main Methods:

• A new algorithm combining quantile regression and an iteratively reweighting strategy was developed.
• The iterative reweighting adjusts residual weights between the fitted baseline and original signals.
• The method does not require user input or prior information like peak detection.

Main Results:

• The proposed algorithm demonstrated speed, flexibility, and robustness across various analytical signal types.
• Performance was validated through comparisons with popular existing baseline correction techniques.
• The method proved effective on both simulated and real-world datasets.

Conclusions:

• The novel quantile regression and iteratively reweighting algorithm provides an effective, automated solution for baseline correction.
• This approach enhances the reliability of data analysis from complex analytical signals.
• The algorithm is user-friendly and suitable for diverse applications in analytical chemistry.