Journal of Chemometrics最新文献

Multivariate extensions of repeated measures linear mixed models, such as repeated measures ANOVA simultaneous component analysis (RM-ASCA+) and linear mixed model-principal component analysis (LiMM-PCA), can be used for analyzing longitudinal studies with multivariate outcomes. However, there are no gold standards to assess the statistical validation of the observed effects of such models. Using real and simulated data, we here perform an empirical comparison of different strategies for assessing statistical significance in these frameworks: permutation tests, the global log-likelihood ratio (GLLR) test, and nonparametric bootstrap confidence intervals for the estimated multivariate effects. Power curves were used to examine the statistical power of the different tests in detecting time–treatment interactions with varying effect sizes. Our results show that both the permutation tests and the GLLR test can be used to statistically test the presence of a time–treatment interaction effect for multivariate data; however, the GLLR approach will be sensitive to the number of included principal components in LiMM-PCA. The bootstrap confidence interval approach generally shows good statistical power but has inflated Type 1 error rates under certain conditions. This makes it unsuitable for the purpose of hypothesis testing in its present implementation, although it may still be useful for exploratory purposes. Overall, our results show that the power of the tests for assessing multivariate effects in longitudinal studies is dependent on characteristics of the dataset, and it is important to be aware of the strengths and weaknesses of the different validation procedures.

An intelligent and accurate modeling method is proposed combining near-infrared (NIR) spectroscopy for measuring organic matter content in soil samples. The proposed method uses Monte Carlo (MC) random sampling in the training set, where subsets were randomly selected from the samples and further selected using the butterfly optimization algorithm (BOA) to construct partial least squares (PLS) submodels, named MC-BOA-PLS. Ultimately, the final prediction was obtained by averaging the predictions of these submodels. The parameters of the MC-BOA-PLS model were optimized, including the iteration number of BOA, the number of butterflies, and the number of PLS submodels. Results show that MC-BOA-PLS exhibited superior predictive performance to predict organic matter content in soil compared with PLS and BOA-PLS.

Wavelength selection (WS) is an effective means to address the presence of many uncorrelated and collinear variables in high-dimensional spectral data that seriously influence the modeling accuracy and efficiency. Aiming to address too many wavelength variables selected by particle swarm optimization algorithm (PSO) and its premature convergence, this paper proposes a novel spectral WS approach—iPSOSA—based on the improved PSO idea and simulated annealing algorithms (SA) strategy. iPSOSA applies the velocity and position update ideas of PSO to the guided shift evolution process of the binary bits with the value of “1” in the particle and integrates with the perturbation strategy of the SA Metropolis acceptance criterion. It effectively solves the premature convergence of PSO and overcomes the low efficiency of the SA evolution, which has high efficiency in WS. By evaluating the modeling performance of different intelligent WS methods using two public spectral datasets from soil and maize, it was found that the iPSOSA outperforms the full-spectrum and other three comparative algorithms. The best iPSOSA partial least squares regression models for soil organic matter and maize moisture contents have excellent regression performance, with the validation set's coefficient of determination higher than 0.98, relative root mean squared error lower than 1.50%, and residual predictive deviation higher than 8.00. iPSOSA presents better comprehensive performance in WS than traditional intelligent algorithms in terms of modeling performance, variable dimensionality, and searching efficiency, providing a new solution for obtaining high correlation wavelength variables in the spectral modeling process.

The comparison of graphs using various types of quantitative structural similarity or distance measures has an important place in many scientific disciplines. Two of these are cheminformatics and chemical graph theory, in which the structural similarity or distance measures between molecular graphs are analyzed by calculating the Jaccard/Tanimoto index based on molecular fingerprints. A novel method is proposed to measure the structural similarity or distance for molecular and arbitrary graphs. This method calculates the Jaccard/Tanimoto index based on a collection of topological indices embedded in the entries of a vector. We statistically compare the proposed method with the method for calculating the Jaccard/Tanimoto indices based on five different molecular fingerprints on alkane and cycloalkane isomers. Furthermore, to explore how the method works on non-molecular graphs, we statistically analyze it on the set of all connected graphs with seven vertices. The Jaccard/Tanimoto index values produced by the proposed method cover the value domain. In addition, it provides a discrete similarity distribution with the clustering, which makes the differences clear and provides convenience for comparison. Two outstanding features of the proposed method are its applicability to arbitrary graphs and the computational complexity of the algorithm used in the method is polynomial over the number of graphs and the number of vertices and edges of the graphs.

This study presents a novel Monte Carlo–based methodology for estimating classification uncertainty in chemometric models by propagating spectral measurement noise. Unlike traditional approaches that treat classification as deterministic, this method simulates realistic noise structures, both independent and correlated, captured from multiple spectrum measurements to quantify sample-specific uncertainty. The technique is applicable to both linear and non-linear models, including partial least squares discriminant analysis (PLS-DA) and various support vector machine (SVM) kernels. The methodology was validated using three datasets: synthetic 2D simulations for controlled model geometry, X-ray fluorescence (XRF) spectra from colored glass rods, and laser-induced breakdown spectroscopy (LIBS) data from Dalbergia wood species. Results revealed that uncertainty increases with spectral similarity and perpendicular alignment between noise structures and decision boundaries. In real-world applications, classification metrics alone proved insufficient to assess model reliability. The inclusion of uncertainty intervals enabled identification of ambiguous predictions even in cases of perfect classification accuracy. This work advances chemometric analysis by linking measurement uncertainty to classification outcomes, offering a robust framework for decision-making in high-stakes analytical contexts.

The effectiveness of artificial neural networks, which were key technologies in artificial intelligence, greatly depends on the quality of the input data. Data cleaning, a crucial component of data preprocessing, played a vital role in enhancing the accuracy, robustness, and generalization capabilities of neural network models. In this study, a Feedback-Driven Iterative Cleaning (FDIC) framework, guided by model performance, was developed and applied to the study of droplet size prediction models for nanocellulose-stabilized Pickering emulsion systems. After randomly removing between 1% and 40% of the data, an artificial neural network model was established using CNC particle size (X1), CNC concentration (X2), and the oil–water volume ratio of CNC to oil-phase monomer (X3) as input variables, with emulsion droplet size (Y) as the quantitative index. The model's accuracy was evaluated after data removal using the coefficient of determination (R²), mean squared error (MSE), and mean absolute scaling error (MASE). The main finding was that targeted removal of a small portion of the data significantly improved the predictive power of the model. Specifically, removing 5% of the dataset results in optimal performance, with R² improving from 0.5307 without cleaning to 0.7258, with an MSE of 183.4917, and MASE of 0.4060. This result suggested a significant and quantifiable improvement in the accuracy of the model through our iterative cleaning process. The study revealed a nonlinear relationship between the number of iterations and the model's generalization ability. This finding offered a novel methodological tool for data governance in the smart era and demonstrates significant value in dynamic environments.

In order to achieve nondestructive analysis and identification of the relative formation time of paper evidence and to solve the difficulties in document authenticity identification in the field of forensic science, this study selected three-dimensional fluorescence spectroscopy data of paper evidence of the same brand and model collected in the same storage environment within the last decade (2012–2023). After preprocessing steps like eliminating scattering, smoothing noise and principal component analysis (PCA), machine learning algorithms such as K-nearest neighbor (KNN) and linear discriminant analysis (LDA) were employed to classify and predict specific feature bands. The accuracy of KNN and LDA was 94.5% and 98.9%, respectively. Furthermore, relative formation time prediction was conducted for paper samples by LDA in the sample library, achieving an accuracy rate of 98.0%. Finally, the established model was successfully applied to analyze an actual case involving suspected “forged official documents.” It accurately determined the relative formation time of the forged paper, and the analysis results were consistent with the suspect's confession.