Framework for multivariate selectivity analysis, part I: theoretical and practical merits.

Area of Science:

• Analytical Chemistry
• Chemometrics

Background:

• Existing multivariate selectivity definitions, particularly Net Analyte Signal (NAS) based, face challenges with inverse calibration methods.
• Classical figures of merit are often misinterpreted when applied to inverse models due to differing properties.

Purpose of the Study:

• To propose a theoretically consistent and experimentally measurable framework for multivariate selectivity.
• To develop a selectivity definition applicable across various calibration methods, including inverse least-squares regression.
• To provide a method consistent with traditional selectivity definitions and interference testing guidelines.

Main Methods:

• Developed a theoretical framework defining selectivity as a function of analyte concentration change caused by interferant concentration change.
• Proposed experimental measurement through controlled selectivity experiments or analysis of existing sample data.
• Illustrated the framework using simulated data and will apply it to near-infrared datasets in Part II.

Main Results:

• Introduced a novel selectivity framework that is theoretically sound and practically measurable.
• Demonstrated the framework's applicability irrespective of the calibration method employed.
• Established a selectivity definition focused on specific analyte-interferant pairs, unlike NAS-based approaches.

Conclusions:

• The proposed selectivity framework offers a unified and robust approach for chemometric analysis.
• This method enhances the interpretation of calibration models, especially inverse methods.
• The framework provides a practical tool for assessing and understanding multivariate selectivity in analytical measurements.