ORIGINAL ARTICLE Discrete Wavelet Transform (DWT) Assisted Partial Least Square (PLS) Analysis of Excitation-Emission Matrix Fluorescence (EEMF) Spectroscopic Data Sets: Improving the Quantification Accuracy of EEMF Technique Keshav Kumar 1 Received: 28 September 2018 /Accepted: 19 November 2018 # Springer Science+Business Media, LLC, part of Springer Nature 2018 Abstract In the present work, it is shown that quantitative estimation efficiency of the partial least square (PLS) calibration model can be significantly improved by pre-processing the EEMF with discrete wavelet transform (DWT) analysis. The application of DWT essentially reduces the volume of data sets retaining all the analytically relevant information that subsequently helps in estab- lishing a better correlation between the spectral and concentration data matrices. The utility of the proposed approach is successfully validated by analyzing the dilute aqueous mixtures of four fluorophores having significant spectral overlap with each other. The analytical procedure developed in the present study could be useful for analyzing the environmental, agricultural, and biological samples containing the fluorescent molecules at low concentration levels. Keywords Excitation-emission matrix fluorescence . Partial least square analysis . Wavelet analysis . Discrete wavelet analysis . Fluorophores Introduction Over the years, the technological developments have provided the desired sensitivity and selectivity to the modern fluorimeters that essentiality has made it possible to analyze the mixtures of fluorophores in a simple and swift manner. Excitation-emission matrix fluorescence (EEMF) spectroscopy is the most com- monly used fluorescence technique [1–5]. EEMF has been suc- cessfully used for analyzing the samples of biological, clinical, pharmaceutical, petrochemicals and environmental origin [1–5]. The popularity of EEMF technique as an analytical tool can be attributed to the following two reason (i) most of the fluorimeters have the required software that allows the acquisi- tion of the EEMF spectrum without much inputs form the user and (ii) it allows capturing the fluorescence response of all the fluorophores in a single spectrum [1–5]. EEMF spectrum es- sentially consists of emission and excitation spectra acquired at different excitation and emission wavelengths, respectively. EEMF technique serve as a useful tool to Bfingerprint^ the fluorescent molecules in a reliable manner. It is true that with the recent advancement in the spectro- fluorometric technology the acquisition of EEMF spectral profiles have become much easier. However, the analyses of large volume of EEMF data sets remain a computational and interpretive challenge. To make meaningful interpretation in a swift manner, over the years several chemometric approaches have been successfully used. Partial least square (PLS) algo- rithm [6–10] is one of the most commonly used chemometric technique that has been used with EEMF to develop a reliable and robust calibration model for quantifying the analytes with- out involving any pre-separation step. The PLS algorithm search for the significant factors that describes the maximum variance of the spectral datasets and maximizes the correlation between the spectral and concentration information of the fluorophores. The PLS modeling requires that set of (i) spec- tral variables and (ii) latent variables be optimized. The opti- mizations of these two parameters are interrelated, as a result, it becomes a time consuming and conceptually and computa- tionally challenging task. The optimization is mainly targeted towards minimizing the error of prediction for the * Keshav Kumar keshavkumar29@gmail.com 1 Institute for Wine analysis and Beverage Research, Hochschule Geisenheim University, 65366 Geisenheim, Germany Journal of Fluorescence https://doi.org/10.1007/s10895-018-2327-z