Giussani, Barbara; Gorla, Giulia; Ezenarro, Jokin; Riu, Jordi; Boque, Ricard (2024). Navigating the complexity: Managing multivariate error and uncertainties in spectroscopic data modelling. Trac-Trends In Analytical Chemistry, 181(), 118051-. DOI: 10.1016/j.trac.2024.118051
Papper original source:
Trac-Trends In Analytical Chemistry. 181 118051-
Abstract:
Spectroscopy and chemometrics, supported by computer science, have yielded promising outcomes, as evidenced by trends observed in literature searches. However, while researchers meticulously construct chemometric models for exploratory, quantitation and classification purposes, the investigation of data quality, particularly error analysis, remains less frequent. Understanding and quantifying measurement errors is crucial for robust spectroscopic modeling and uncertainty estimation. By unraveling complexities related to multivariate errors and uncertainties in spectroscopic data, the scientific community is empowered to extract reliable information from spectroscopic analyses, paving the way for enhanced analytical practices. This review underscores the necessity for the scientific community to integrate error analysis and uncertainty estimation into multivariate analysis methods, offering tailored solutions for diverse data types and analysis objectives.
Spectroscopy and chemometrics, supported by computer science, have yielded promising outcomes, as evidenced by trends observed in literature searches. However, while researchers meticulously construct chemometric models for exploratory, quantitation and classification purposes, the investigation of data quality, particularly error analysis, remains less frequent. Understanding and quantifying measurement errors is crucial for robust spectroscopic modeling and uncertainty estimation. By unraveling complexities related to multivariate errors and uncertainties in spectroscopic data, the scientific community is empowered to extract reliable information from spectroscopic analyses, paving the way for enhanced analytical practices. This review underscores the necessity for the scientific community to integrate error analysis and uncertainty estimation into multivariate analysis methods, offering tailored solutions for diverse data types and analysis objectives.