When buying apples, there are several characteristics that are important to the consumer. These include taxonomic variety, which influences the desired flavor and the allergenicity, and, ...increasingly, whether the cultivation is regional and organic. In this study, we analyzed 217 apple samples with 1H NMR spectroscopy and verified the authentication of the apples in terms of country of origin, local origin within Germany, taxonomic variety and whether they were organically produced. The obtained random forest models showed classification accuracies ranging from about 73 % for the discrimination of seven taxonomic varieties to 88.5 % for the separation of German and non-German samples. The results demonstrate that 1H NMR spectroscopy is promising for comprehensive authentication of apples. This is achieved by using different parts of the NMR fingerprint depending on the desired outcome, which was confirmed by the application of variable selection methods and the comparison of the selected variables.
•217 apple samples were analyzed with 1H NMR spectroscopy.•NMR signals were assigned to known apple metabolites.•Random forest was applied for the differentiation of apples according to five different authentication issues.•Relevant signals were selected by random forest variable selection.•Exclusive spectral regions for each authentication issue could be identified.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Over the past two decades, comparative sequence analysis using codon-substitution models has been honed into a powerful and popular approach for detecting signatures of natural selection from ...molecular data. A substantial body of work has focused on developing a class of "branch-site" models which permit selective pressures on sequences, quantified by the ω ratio, to vary among both codon sites and individual branches in the phylogeny. We develop and present a method in this class, adaptive branch-site random effects likelihood (aBSREL), whose key innovation is variable parametric complexity chosen with an information theoretic criterion. By applying models of different complexity to different branches in the phylogeny, aBSREL delivers statistical performance matching or exceeding best-in-class existing approaches, while running an order of magnitude faster. Based on simulated data analysis, we offer guidelines for what extent and strength of diversifying positive selection can be detected reliably and suggest that there is a natural limit on the optimal parametric complexity for "branch-site" models. An aBSREL analysis of 8,893 Euteleostomes gene alignments demonstrates that over 80% of branches in typical gene phylogenies can be adequately modeled with a single ω ratio model, that is, current models are unnecessarily complicated. However, there are a relatively small number of key branches, whose identities are derived from the data using a model selection procedure, for which it is essential to accurately model evolutionary complexity.
Land cover mapping in complex environments can be challenging due to their landscape heterogeneity. With the increasing availability of various open-access remotely sensed datasets, more images ...acquired by different sensors and on different dates tend to be used to improve land cover classification accuracy. Selecting an appropriate feature domain with the best landscape separability is therefore crucial in meeting the requirement of computational efficiency and model interpretability. Variable selection is widely used in pattern recognition to enhance model parsimony. This study focused on the variable selection process and proposed a series of methods to select the optimal feature domain to improve land cover classification in a complex urbanized coastal area. Two decision tree models (CART-Classification and Regression Tree and CIT-Conditional Inference Tree) and five variable importance measures (GINI, PVIM-Permutated Variable Importance Measure, MD- Minimum Depth, IPM-Intervention of Prediction Measure, and CPVIM-Conditional Permutation Variable Importance Measure) based on random forests were considered. Variable importance measures were applied to a set of spectral, spatial and temporal features derived from medium-resolution satellite images. Backward elimination methods were used to select the optimal feature subset. It is found that compared to the traditional band-only model, the variable selection process can significantly improve the model parsimony and computational efficiency. The CPVIM based on CIT decision tree model was more reliable in selecting relevant features regardless their correlations, but CART tended to generate higher classification accuracy. Therefore, the combination of the CART model and the ranking from the CPVIM variable measure is recommended to achieve higher classification accuracy and better data interpretability. The novelty of our work is with the insight into the merits of integrating variable selection in the land cover classification process over complex environments.
•Variable selection can significantly improve coastal land cover classification.•The selection of variable importance measures may vary by data types.•Conditional permutated variable importance measure was reliable for correlated data.•Conditional Inference Tree took more time but did not necessarily improve accuracy.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
A Bayesian bi-level variable selection method (BAGB: Bayesian Analysis of Group Bridge) is developed for regularized regression and classification. This new development is motivated by grouped data, ...where generic variables can be divided into multiple groups, with variables in the same group being mechanistically related or statistically correlated. As an alternative to frequentist group variable selection methods, BAGB incorporates structural information among predictors through a group-wise shrinkage prior. Posterior computation proceeds via an efficient MCMC algorithm. In addition to the usual ease-of-interpretation of hierarchical linear models, the Bayesian formulation produces valid standard errors, a feature that is notably absent in the frequentist framework. Empirical evidence of the attractiveness of the method is illustrated by extensive Monte Carlo simulations and real data analysis. Finally, several extensions of this new approach are presented, providing a unified framework for bi-level variable selection in general models with flexible penalties.
•A Bayesian bi-level variable selection method is developed for linear models.•A data-augmentation scheme is introduced based on a novel mixture representation.•Posterior inference is carried out via a computationally efficient MCMC algorithm.•Benchmark results indicate superior performance compared to published methods.•Simple extensions are discussed for general models with flexible penalties.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK, ZRSKP
We develop a new primal-dual witness proof framework that may be used to establish variable selection consistency and ℓ∞-bounds for sparse regression problems, even when the loss function and ...regularizer are nonconvex. We use this method to prove two theorems concerning support recovery and ℓ∞-guarantees for a regression estimator in a general setting. Notably, our theory applies to all potential stationary points of the objective and certifies that the stationary point is unique under mild conditions. Our results provide a strong theoretical justification for the use of nonconvex regularization: For certain nonconvex regularizers with vanishing derivative away from the origin, any stationary point can be used to recover the support without requiring the typical incoherence conditions present in ℓ1-based methods. We also derive corollaries illustrating the implications of our theorems for composite objective functions involving losses such as least squares, nonconvex modified least squares for errors-in-variables linear regression, the negative log likelihood for generalized linear models and the graphical Lasso. We conclude with empirical studies that corroborate our theoretical predictions.
Full text
Available for:
BFBNIB, INZLJ, NMLJ, NUK, PNG, SAZU, UL, UM, UPUK, ZRSKP
Hybrid variable selection method based on V-WSP and variable importance measurement (VIM) coupled with random forest (RF) for quantitative analysis of methanol content in methanol gasoline.
Display ...omitted
•A hybrid variable selection strategy based on V-WSP-VIM coupled with RF was constructed.•OOB error estimation and 10-flod cross validation were applied to improve RF model.•The V-WSP-VIM method can achieve better prediction performance than several other methods.
With the trend of portable and miniaturization, Raman spectrometer requires more advanced analytical methods providing more rapid and accurate analysis performance for in-situ analysis. In this work, a hybrid variable selection method based on V-WSP and variable importance measurement (VIM) coupled with random forest (RF) was used to improve the quantitative analysis performance of portable laser Raman instruments for quantitative analysis of methanol content in methanol gasoline. First, five preprocessing methods were applied to reduce the infection information in the raw spectra, respectively. Based on the spectra data processed by multivariate scattering correction (MSC), V-WSP was employed to filter the infection or redundant information in Raman spectroscopy, and 579 variables were obtained when the correlation threshold is 0.9600. Then, the variables were further eliminated by VIM. Finally, 43 variables were obtained by the V-WSP-VIM method. In data processing, out of bag (OOB) error estimation and 10-flod cross validation (CV) were applied to optimize the parameters of preprocessing methods, V-WSP, VIM and RF model. The results fully demonstrated that compared with the RF model based on raw spectra, the RF model based on V-WSP-VIM method can achieve a better prediction performance for the quantitative analysis of methanol content in methanol-gasoline, with the coefficients of determination of cross-validation (R2CV) improving from 0.9100 to 0.9662, the root mean square error of cross-validation (RMSECV) reducing from 0.0572 to 0.0365%, the coefficients of determination of prediction set (R2P) improving from 0.9214 to 0.9407, the root mean square error of prediction set (RMSEP) reducing from 0.0420 to 0.0382%, the variables reducing from 1044 to 43 and the modeling time reducing from 72.94 to 6.41 s. The results indicates that V-WSP-VIM coupled with RF is an effective method to improve the performance of portable laser Raman spectrometer for quantitative analysis of methanol content in methanol gasoline.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
This study utilized least angle regression (LARS) with the least absolute shrinkage and selection operator (LASSO) to select important wavelengths for rapid quantification of vitamin C (Vc) and ...reducing sugar (RS) in sweetpotato roots (SPR) using hyperspectral imaging (900–1700 nm). Nine wavelengths strongly correlated with Vc levels and twelve with RS were identified, achieving good predictions with partial least squares (PLS) regression (Vc: rP = 0.9704, RMSEP = 1.0098 mg/100 g; RS: rP = 0.9641, RMSEP = 0.2725 g/100 g). Validation with an independent sample set (n = 35) showed minimal deviation between predicted and actual values (Vc: 1.179–1.211 mg/100 g; RS: 0.316–0.324 g/100 g). Explainable AI and SHapley Additive exPlanations (SHAP) values were used to interpret the selected wavelengths. Chemical maps visually analyzed Vc and RS distribution in different SPR samples. This method effectively estimates Vc and RS levels in SPR, potentially aiding SPR quality assessment during post-harvest marketing and storage.
•Spectra were used to assess vitamin C and reducing sugar of sweetpotato roots.•LARS-LASSO algorithm was applied to select most informative wavelengths.•Vitamin C and reducing sugar contents in sweet potato roots were well predicted.•Spatial differences of vitamin C and reducing sugar in samples were visualized.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
58.
What are the true clusters? Hennig, Christian
Pattern recognition letters,
10/2015, Volume:
64
Journal Article
Peer reviewed
Open access
•A constructivist view of reality and science is sketched.•Context- and aim-dependent characteristics of clusterings are listed.•Formal approaches to define true clusters are presented.•Researchers ...need to communicate their cluster concept transparently.•Comparisons should show how different methods are good for different aims.
Constructivist philosophy and Hasok Chang’s active scientific realism are used to argue that the idea of “truth” in cluster analysis depends on the context and the clustering aims. Different characteristics of clusterings are required in different situations. Researchers should be explicit about on what requirements and what idea of “true clusters” their research is based, because clustering becomes scientific not through uniqueness but through transparent and open communication. The idea of “natural kinds” is a human construct, but it highlights the human experience that the reality outside the observer’s control seems to make certain distinctions between categories inevitable. Various desirable characteristics of clusterings and various approaches to define a context-dependent truth are listed, and I discuss what impact these ideas can have on the comparison of clustering methods, and the choice of a clustering methods and related decisions in practice.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK
AbstractIn the dynamic landscape of banking institutions, acquiring accurate and timely information regarding customers’ incomes is crucial for effectively managing financial product offerings. To ...meet this demand, these institutions construct predictive models using numerous features, with only a subset contributing to capturing income variability. In this study, we propose a methodology for predicting monthly incomes by employing an XGBoost model with a reduced number of features. Feature reduction is accomplished through the implementation of Boruta and BorutaSHAP, ensuring that no predictive power is lost throughout the process. To enhance the transparency of the model’s predictions, we used the Shapley Additive Explanations (SHAP) method. The dataset used was provided by an anonymous bank from Chile, consisting of 10,000 records, 426 features, and a substantial proportion of missing values. The results demonstrate that the combination of feature selection methods and the XGBoost algorithm enables the development of a more concise model that maintains predictive performance. By leveraging the SHAP method, financial institutions can consistently identify and track influential features, thereby reducing complexity and training time without compromising predictive power. This research offers valuable contributions to financial institutions, as they can adopt our methodology to consistently identify and track the most influential features.