The researchers recruited 152 people with etomidate use disorder and 67 healthy controls. Participants completed standardized assessments measuring craving, impulsivity, inhibitory control and habitual behavior. The etomidate use disorder group had significantly higher craving and impulsivity scores and poorer performance on the stop-signal and reversal-learning tasks. The available abstract does not provide further details about the procedures, effect sizes or the specific performance measures used in those tasks.

Resting-state EEG data were examined using microstate analysis to characterize the temporal organization of large-scale brain activity. Compared with controls, the etomidate use disorder group had significantly longer durations of microstates B, C and D and a reduced occurrence of microstate A. The researchers also built a multidimensional feature-based support vector machine model using EEG features. In distinguishing participants with etomidate use disorder from healthy controls, the model achieved an area under the receiver operating characteristic curve, or AUC, of 0.777.

The abstract describes these findings as cross-sectional clinical and neurophysiological features associated with etomidate use disorder, rather than evidence that the disorder caused the observed differences. Because the study compared groups at one point in time, it cannot establish whether the behavioral or EEG differences preceded, resulted from or developed alongside the disorder. The AUC represents discrimination performance in the study’s participant sample, not a validated clinical diagnostic test. The available evidence is limited to the published abstract, which does not report detailed recruitment criteria, participant characteristics beyond age matching, the complete set of EEG features, validation procedures or broader generalizability.

Subject of Research: Clinical and EEG microstate features associated with etomidate use disorder

Article Title: Impulsivity, cognitive inflexibility, and brain network dysregulation in etomidate use disorder: A clinical and EEG microstate study

Article References: Liao, Z., Huang, Q., Chen, X., Le, J., Wang, X., Wang, X., Tang, Y., & Shen, H. (2026). Impulsivity, cognitive inflexibility, and brain network dysregulation in etomidate use disorder: A clinical and EEG microstate study. Scientific Reports. https://doi.org/10.1038/s41598-026-75726-3

Image Credits: AI Generated

DOI: 10.1038/s41598-026-75726-3

Keywords: etomidate use disorder, c Craving, impulsivity, inhibitory control, EEG microstates, brain activity, support vector machine, cross-sectional study, cognitive, inflexibility, brain, network

The study used a publicly available Kaggle dataset containing images divided among four classes: Black Spot, Canker, Fresh and Greening. It evaluated eight fixed feature extractors—seven convolutional neural networks and a ViT-S/16 Vision Transformer—together with four classical classifiers: SVM, k-nearest neighbors, random forest and XGBoost. All image models were evaluated using the same dataset splits, preprocessing and metrics under 5 × 5 repeated stratified cross-validation, comprising five repeats of five folds with an 80:20 class-stratified train-test ratio in each fold.

To prevent information from held-out images entering the dimensionality-reduction stage, researchers fitted PCA separately on the training-fold features in every iteration and then applied the resulting transformation to the corresponding test fold without refitting it. The researchers used a target of retaining 98% of feature variance. A sensitivity analysis found that mean accuracy plateaued across the 95%–98% range and did not improve at 99%, supporting the selected threshold.

Across the 32 backbone-classifier combinations, SVM was consistently the strongest or nearly strongest classifier after PCA reduction. For ViT features, SVM significantly outperformed k-nearest neighbors, random forest and XGBoost, with all reported comparisons having p < 0.001. Paired tests across the 25 matched partitions also found that the ViT + PCA + SVM combination significantly outperformed DenseNet201 + PCA + SVM (p < 0.001) and EfficientNetB0 + PCA + SVM (p = 0.004).

The class-level results for the proposed ViT pipeline were not limited to the two easiest categories. Aggregated over the 25 held-out folds, F1-scores were 97.70% for Black Spot and 97.63% for Canker, compared with 99.97% for Fresh and 100.00% for Greening. The two visually similar disease classes therefore had somewhat lower, though still high, F1-scores than the other categories.

Ablation tests indicated that PCA’s main measured contribution was feature reduction rather than an increase in accuracy. For ViT, retaining 98% of variance required an average of 170 of 384 components, a reduction of about 55.7%. ViT + PCA + SVM was statistically indistinguishable from ViT features without PCA (99.01% ± 0.62%; p = 0.16) and from SVM applied to the full 384-dimensional feature space (99.12% ± 0.69%; p ≈ 1.0). The corresponding ablation for DenseNet201 produced a larger reduction, about 76.6%, while the complete pipeline did not significantly differ from unreduced DenseNet201 features (p = 0.25).

Explainability analyses used occlusion sensitivity and SHAP for the ViT configuration, while DenseNet201 was assessed with Grad-CAM and SHAP because the study states that Grad-CAM is not directly applicable to transformer attention. ViT occlusion maps showed high sensitivity in localized fruit-surface regions for Fresh and Canker, distributed sensitivity across the lower half of the fruit for Greening, and sensitivity around visible dark spots for Black Spot. SHAP summaries showed contributions distributed across multiple principal components, with both positive and negative values. These representative examples indicated reliance on fruit-surface evidence, although some ViT maps also showed sensitivity near image borders.

The computational benchmark also distinguished accuracy from model size. ViT had 21.67 million parameters and an approximate size of 82.6 MB, while DenseNet201 had 18.32 million parameters and an approximate size of 69.9 MB. The measured mean inference time was 99.07 ms for ViT and 118.66 ms for DenseNet201. EfficientNetB0 and MobileNetV2 were substantially smaller, with 4.05 million and 2.26 million parameters, respectively.

The findings are bounded by the dataset’s modest size and single public source. The 1,090 field or orchard photographs were whole intact fruit, but their repository did not provide original per-image acquisition details such as camera device, precise orchard location or controlled-versus-field capture conditions. The images also may not represent the full variation in cultivars, growing regions, lighting and sensors encountered elsewhere. The authors therefore characterized the reported accuracies as indicative rather than conclusive and identified validation on independently collected, multisource imagery as necessary future work.

Subject of Research: Experimental benchmarking of PCA and SVM classifiers with deep image features for four-class orange fruit disease classification

Article Title: A PCA-based deep feature optimization framework for explainable orange fruit disease classification

Article References: Hingmire, A., Golande, A., Bhutnal, V., Pande, S. D., & Pande, T. (2026). A PCA-based deep feature optimization framework for explainable orange fruit disease classification. BMC Plant Biology, 26(1), Article 1746. https://doi.org/10.1186/s12870-026-09984-8

Image Credits: AI Generated

DOI: 10.1186/s12870-026-09984-8

Keywords: orange diseases, image classification, Vision Transformer, principal component analysis, support vector machine, explainable AI, cross-validation, DeepNet201, computer vision, fruit disease dataset, PCA-based, deep

The core idea is elegant. In an ideal OTU, every sequence would sit at a dissimilarity smaller than the barcoding gap, the distance threshold that separates sequences of one taxon from those of another, with every read linked to every other read. Real OTUs deviate from this theoretical pattern, and the authors argue that the deviations are not random noise but carry structure of their own. They identify two characteristic ways in which an OTU can go wrong. The first is the composed OTU, an artificial merger of several true units, typically produced by the well-known chaining effect of single linkage clustering, in which sequences A and B are close, B and C are close, but A and C are far beyond the barcoding gap. The second is the OTU with noise, in which a handful of satellite sequences hang loosely off a densely connected core without forming a compact subgroup of their own.

To detect composed OTUs, the method relies on a single statistic computed directly from the array of pairwise dissimilarities within each OTU. The researchers build a graph in which each read is a node and an edge connects two reads whenever their dissimilarity falls below the barcoding gap. They then compute theta, the ratio of missing edges to all possible edges in that graph. For a single OTU, nearly all dissimilarities lie below the gap, so theta is small. For a composed OTU containing two balanced entities, the inter-entity dissimilarities are mostly above the gap, and theta approaches one half. Sorting the OTUs of a sample then becomes a matter of building a smoothed histogram of theta values with a Gaussian kernel and placing the critical threshold at the first local minimum between the first and second modes of the estimated density. Crucially, unlike the breaking phase of the SWARM algorithm, this procedure requires no arbitrarily fixed parameter.

Once composed OTUs have been separated from single ones, a second, supervised step hunts for noise among the single OTUs. The authors fit a two-block Stochastic Block Model to each dissimilarity array, a statistical framework that clusters sequences according to their shared patterns of connectivity and has the unusual plasticity of allowing blocks to be either assortative or dissortative. If an OTU is clean, both blocks are assortative and the diagonal elements of the inferred connectivity matrix are small. If noise reads are present, they are grouped into a dissortative block with large intra-block and inter-block dissimilarities. The two discriminating features, the mean inter-block dissimilarity and the largest mean intra-block dissimilarity, feed a linear Support Vector Machine trained on OTUs labelled by a degree-based expert classifier, which examines the percentage of gap-level connections each read maintains with the rest of the OTU.

Validation came from a rich dataset of diatoms collected in Arcachon Bay on the French Atlantic coast, made available through the Malabar project. Thirty-two samples were balanced across four seasons, four locations and two water columns, pelagic at high tide and benthic, yielding between 19,000 and 36,000 reads per sample of a 312 base pair region of the rbcL chloroplastic marker. Pairwise dissimilarities were computed with Smith-Waterman local alignment scores on a computing cluster, and OTUs were extracted as connected components of the gap-level graph, keeping only units with fifteen reads or more. For the eight samples of the Teychan location, an expert classification was built by visual inspection of heatmaps and community-structured graphs of 654 dissimilarity arrays, and the agreement with the automatic method was striking: only twelve false negatives among 104 expertly composed OTUs and nine false positives among 527 single ones.

Applied to the full dataset of 2,529 OTUs, the method revealed strong statistical links between OTU type and OTU size. A Wilcoxon Mann-Whitney test returned a p-value of roughly 1.5 times ten to the power of minus fifty-five for the composed-versus-single distinction, confirming that composed OTUs are systematically the largest ones. Among single OTUs, the test for noise returned a p-value of about 3.7 times ten to the power of minus twenty-two, showing that noisy units are also significantly larger than clean ones. Cross-validation of the Support Vector Machine on the training set achieved an Area Under Curve of 0.951, and on the 24 test samples the classifier caught every OTU with noise, although it recovered only about 64 percent of the OTUs without noise, a limitation the authors attribute to the gradient of noise intensity that the two SBM features trace simultaneously.

Taxonomic annotation provided an independent check. Among 180 fully annotated OTUs, about 85 percent were monospecific, and almost all single OTUs with noise, 153 of them, were monospecific as well, suggesting that satellite reads often belong to the same species as the core rather than to a different taxon. Only eight fully annotated OTUs were typed as composed, and they fell into three profiles: genuine multi-species mergers with clear two-block structure, monospecific units with loosely connected reads, and monospecific units whose dissimilarity arrays nevertheless split into two blocks, possibly reflecting structured intraspecific molecular diversity. One artefact also surfaced: a single reference sequence of the diatom Rhizosolenia fallax appeared repeatedly across OTUs of several genera and was set aside as a mapping anomaly.

Perhaps the most provocative result concerns the environment itself. The proportions of composed, noisy and clean OTUs varied from sample to sample, and a Wilcoxon rank test on the fractions of composed OTUs in benthic versus pelagic samples returned a p-value of 1.6 times ten to the power of minus four. Benthic samples carried a mean fraction of 0.19 composed OTUs against 0.10 in pelagic ones, strong evidence that the molecular diversity of the benthic diatom flora is structured differently, with species distances globally smaller than in the pelagic realm. The authors hypothesise that habitat may influence the alternation of sexual and asexual reproduction and thereby shape patterns of DNA diversity, though the chloroplastic rbcL marker makes any such influence indirect. No difference emerged for the fraction of noisy OTUs among single units.

The practical guidance that follows is a gradient of trust. Cliques, in which every read connects to every other, are the rarest and most reliable units, though only four were found across all 32 samples. Single OTUs without noise, whose SBM parameters stay below eight, below the barcoding gap of nine, can be used as such for ecological studies. Composed OTUs, which are very likely spurious mergers reflecting the inadequacy of a single barcoding gap for the largest clusters, should be post-treated with classical clustering or community detection tools to be reshaped into smaller, more reliable units. Single OTUs with noise, whose quality remains to be tested across a wider diversity of organisms, sit in between and deserve further scrutiny.

In an era when sequencing pipelines stream millions of reads faster than any expert can inspect them one by one, the appeal of this approach lies in its automation and its transparency. Every step, from the theta statistic to the SBM features and the SVM frontier, is elementary enough to be implemented in Python or R, and the associated code is openly available on the forgemia platform of INRAE. Used as a quality filter within a pipeline, the method promises to strengthen the foundations of molecular ecology, letting researchers know which OTUs can carry the weight of a biodiversity inventory and which ones need reshaping before any ecological conclusion is drawn from them.

Subject of Research: An automatic method for evaluating the quality of Operational Taxonomic Units in metabarcoding data based on dissimilarity arrays

Article Title: Simple approaches for evaluation of OTU quality based on dissimilarity arrays

Article References: Cros, M.-J., Frigerio, J.-M., Peyrard, N., & Franc, A. (2024). Simple approaches for evaluation of OTU quality based on dissimilarity arrays. Metabarcoding and Metagenomics, 8, Article e108649. https://doi.org/10.3897/mbmg.8.108649

Image Credits: AI Generated

DOI: 10.3897/mbmg.8.108649

Keywords: metabarcoding, OTU quality, dissimilarity arrays, diatoms, Arcachon Bay, barcoding gap, Stochastic Block Model, Support Vector Machine, chaining effect, biodiversity, rbcL marker, molecular ecology

The research, published in the journal Discover Sustainability by Wilson Nathan Ndackssi and Modeste Kameni Nematchoua, focused on five Cameroonian cities chosen deliberately to span the country’s climatic spectrum: Bertoua, Douala, Garoua, Maroua, and Yaoundé. This geographic spread matters more than it might first appear. Douala sits on the humid Atlantic coast under persistent cloud cover, while Maroua and Garoua lie in the far north, where the Sahelian climate delivers intense, seasonally concentrated sunshine. Yaoundé and Bertoua occupy intermediate equatorial and tropical humid regimes. A model that performs well across all five sites is therefore tested against the full range of atmospheric conditions a solar engineer is likely to encounter in the region, from coastal haze to dry-season dust.

To train and evaluate their models, the researchers drew on NASA’s POWER database, a publicly accessible archive of meteorological and solar radiation variables derived from satellite observations. The dataset covered 27 years, from 1999 to 2026, giving the algorithms a long enough record to capture seasonal cycles, interannual variability, and the influence of climate phenomena on surface radiation. Eight input variables fed the models: mean air temperature, relative humidity, precipitation, wind speed, wind direction, atmospheric pressure, cloud cover, and the day of the year. The inclusion of cloud cover proved particularly significant, since clouds are the dominant modulator of surface solar radiation in the tropics, capable of cutting daily irradiation by more than half during peak wet seasons.

Four machine learning algorithms were put head to head: Linear Regression, Decision Tree, Random Forest, and Support Vector Machine. Each approach embodies a different philosophy of prediction. Linear Regression assumes a straightforward weighted relationship between the meteorological inputs and irradiation, offering transparency but limited flexibility. A Decision Tree partitions the input space through a cascade of binary questions, capturing nonlinear thresholds such as the sharp radiation drop when cloud cover crosses a certain fraction. The Random Forest builds hundreds of Decision Trees on random subsets of the data and averages their outputs, a technique known as ensemble learning that suppresses the overfitting to which single trees are prone. The Support Vector Machine, meanwhile, constructs a decision boundary in a high-dimensional transformed feature space, allowing it to model complex nonlinear relationships without explicitly enumerating them.

The models were evaluated with three standard statistical indicators: the mean absolute error (MAE), which measures the average size of prediction errors in plain units; the root mean square error (RMSE), which penalizes large errors more heavily; and the coefficient of determination (R²), which quantifies the fraction of variance in observed irradiation that the model explains. Across all five cities and all four algorithms, RMSE values ranged from 0.58 to 0.77 kWh/m², MAE values from 0.40 to 0.59 kWh/m², and R² values from 0.52 to 0.69. In practical terms, the best models predicted daily irradiation to within roughly half a kilowatt-hour per square meter on average, a margin meaningful for feasibility studies though not yet tight enough for minute-by-minute grid dispatch.

The ranking of the algorithms was consistent in every city. Random Forest achieved the best predictive performance everywhere, followed by the Support Vector Machine, then the Decision Tree, with Linear Regression trailing last. This ordering tells a coherent story about the physics of the problem. Solar irradiation in the tropics is governed by nonlinear interactions among cloud dynamics, humidity, temperature, and season, and the ensemble averaging of Random Forest is well suited to capturing those interactions while remaining robust to noisy or imperfect inputs. The comparatively modest R² values also signal honest limitations: even the best model explains only about two-thirds of the day-to-day variance, reflecting the inherent chaos of tropical cloud systems and the resolution limits of satellite-derived predictors.

The authors are candid about these constraints, and the study’s title itself foregrounds evidence and limitations alongside its results. The reliance on NASA POWER data means the models inherit the biases of satellite retrieval rather than ground truth, and the researchers note that comparable studies in Cameroon remain few relative to the extensive literature from Asia, Europe, and North Africa. Their work distinguishes itself, they argue, through its complementary geographic coverage, the inclusion of the Support Vector Machine among the tested algorithms, and an in-depth treatment of predictive meteorological variables, cloud cover in particular. The findings position Random Forest as the most suitable algorithm for predicting global daily solar irradiation in Cameroon, and as a preliminary screening tool for assessing solar energy resources across Central Africa.

Why does this matter beyond the modeling community? Cameroon’s solar resources remain largely untapped, even as the country’s electricity demand grows and hydropower, its dominant generation source, becomes vulnerable to increasingly erratic rainfall. Reliable irradiation prediction underpins nearly every stage of solar deployment: it determines the expected energy yield that investors demand before financing a plant, the sizing of battery storage in off-grid systems, and the design of hybrid solar-diesel microgrids for rural communities. A validated, low-cost prediction tool that requires only freely available satellite meteorology could dramatically lower the barrier to preliminary solar resource assessment in a region where installing and maintaining pyranometers, the instruments that measure radiation directly, is often impractical.

The study also carries a broader message for the machine learning and renewable energy communities. Ensemble methods like Random Forest are computationally cheap, interpretable through feature-importance analysis, and trainable on public data, making them realistic options for national energy agencies and small research groups in developing economies, not just well-funded laboratories. At the same time, the moderate R² values are a reminder that prediction is not a substitute for measurement: as solar penetration deepens, ground-truthing campaigns and denser observation networks will be essential to refine the models and to capture local effects that satellites smooth over. The researchers suggest their framework can serve as a first-pass selection tool, with the potential to enrich the scientific literature on solar irradiation modeling in Central Africa, a region notably underrepresented in the global solar forecasting literature.

For a country straddling the equator, the sun is Cameroon’s most dependable energy asset, and the gap between that potential and its exploitation has been shaped as much by information scarcity as by infrastructure. What this study demonstrates is that the information gap is closable with tools already within reach: 27 years of open satellite data, eight meteorological variables, and a forest of decision trees. If the approach is validated further with ground measurements and extended to finer temporal and spatial scales, the same methodology could be replicated across neighboring Central African nations, turning machine learning from a laboratory curiosity into a practical instrument of energy planning for the tropics.

Subject of Research: Machine learning prediction of daily global solar irradiation in the tropical climate zones of Cameroon for renewable energy planning

Article Title: Evidence, limitations and implications for renewable energy planning of daily global solar irradiation in tropical climate zones of Cameroon, predicted by machine learning

Article References: Ndackssi, W. N., & Kameni Nematchoua, M. (2026). Evidence, limitations and implications for renewable energy planning of daily global solar irradiation in tropical climate zones of Cameroon, predicted by machine learning. Discover Sustainability. https://doi.org/10.1007/s43621-026-04807-0

Image Credits: AI Generated

DOI: 10.1007/s43621-026-04807-0

Keywords: solar irradiation, machine learning, Random Forest, Cameroon, renewable energy, NASA POWER database, Support Vector Machine, tropical climate, solar energy planning, Discover Sustainability, cloud cover, Central Africa

The research, led by Sukanta Kumar Tulo of the School of Electronics Engineering at Vellore Institute of Technology in India, together with colleagues at VIT and the Manipal Institute of Technology, takes a deliberately lightweight path. Instead of relying on massive deep neural networks, the team extracted a compact set of morphological imaging biomarkers from segmented lung and mediastinum masks and then fused the most informative ones using wrapper-based feature selection with classical machine learning classifiers. The approach is designed for settings where annotated data, computing power, and expert radiologist time are all in short supply.

The dataset consisted of 200 pediatric chest radiographs—100 bacterial and 100 viral pneumonia cases—drawn from a publicly available collection of images acquired at the Guangzhou Women’s and Children’s Medical Centre. The images, anteroposterior views of children under five years of age, had been reviewed by experts for quality assurance and were collected under ethical committee clearance in accordance with the Declaration of Helsinki. Each image was resampled to 512 by 512 pixels using bicubic interpolation to standardize the inputs for analysis.

Segmentation is the foundation of the pipeline, and the authors built a hybrid model that pairs two level set techniques, each suited to a different anatomical challenge. The lungs were delineated with the Reaction-Diffusion Level Set method, an edge-based approach that incorporates a diffusion term into the level set formulation and thereby eliminates the need for the slow and often unstable re-initialization step required by conventional level set contours. In the study, the method was configured with a Gaussian kernel of standard deviation 1.5, a fixed force of 0.7, a time increment of 0.1, and 800 iterations. The mediastinum—the central compartment of the thoracic cavity bounded laterally by the lungs—was segmented with the Chan-Vese model, a region-based technique that partitions the image according to mean pixel intensities inside and outside a closed contour, with additional regularization terms to ensure boundary smoothness. The mediastinum segmentation ran for 900 iterations, a number chosen empirically.

Once the masks were extracted, the team computed 25 morphological biomarkers per region: 18 geometric descriptors and 7 Hu invariant moments. The geometric set includes area, perimeter, orientation, eccentricity, circularity, equivalent diameter, convex area, solidity, extent, major and minor axis lengths, maximum and minimum Feret diameters and their angles, and bounding box dimensions. Size-related measures such as area and equivalent diameter quantify how large a structure is, while eccentricity and orientation capture elongation and tilt. Compactness indicators such as solidity and circularity describe how closely a region resembles a regular form, and Feret diameters add directional information that is useful for irregular structures. Hu moments, constructed from second- and third-order normalized central moments, are invariant to translation, scale, and rotation, which makes them robust descriptors of shape regardless of how the structure sits in the image.

Statistical analysis with the Wilcoxon rank sum test revealed clear morphological differences between the two pneumonia types. In bacterial cases, the mean lung area was significantly smaller than in viral cases by 14.3 percent, while the mean mediastinum area was 29.6 percent larger, and the mean minor axis length of the mediastinum was 41.9 percent higher, indicating a widened central compartment. The authors attribute the mediastinal enlargement to abnormalities such as bronchial dilatation and mediastinal lymphadenopathy, which enlarge the airways and lymph nodes, as well as pleural effusion, which can shift the mediastinum. In viral images, by contrast, irregular opacities appeared diffusely in both lungs, consistent with interstitial patterns. Roughly 46 percent of bacterial images showed a mediastinum area larger than the lung area, compared with only 21 percent of viral images. Biomarkers including area, minor axis length, eccentricity, circularity, equivalent diameter, minimum Feret diameter, and bounding dimensions were statistically significant for both the left lung and the mediastinum.

To squeeze the most discriminative power out of these measurements, the researchers applied wrapper-based biomarker selection, a strategy that repeatedly trains and tests a learning algorithm with different biomarker combinations to find the subset that yields the best predictive performance. Using forward selection starting from an empty set, they ran the procedure with two algorithms: Linear Discriminant Analysis, which seeks a linear projection that maximizes the ratio of between-class to within-class variance, and Support Vector Machine, which constructs an optimal hyperplane that maximizes the margin between classes. The optimal biomarkers from the left lung, right lung, and mediastinum were then concatenated in series to form fused biomarker sets, which were fed back into LDA and SVM classifiers with a linear kernel and a soft-margin penalty of one, evaluated with five-fold cross-validation.

The results showed a consistent pattern: fusing biomarkers from both regions outperformed using either region alone. With the SVM classifier, raw fused biomarkers achieved an F-measure of 82.9 percent, compared with 75.5 percent for mediastinum biomarkers alone and 81.3 percent for lung biomarkers alone. Wrapper selection improved things further. The LDA-based wrapper combined with the LDA classifier delivered F-measures of 76.5 percent for mediastinum biomarkers, 82.0 percent for lung biomarkers, and a peak of 87.3 percent for the fused set, with an accuracy of 87.0 percent, sensitivity of 89.0 percent, specificity of 85.0 percent, and an area under the ROC curve of 88.3 percent. The authors attribute the superiority of the LDA wrapper to its efficient dimensionality reduction, which helps it select biomarkers that generalize well.

Compared with deep learning studies on the same publicly available dataset, the biomarker fusion approach holds its own while demanding far less. The original Kermany benchmark using the Inception network reported 90.7 percent accuracy, and later deep models such as MD-Conv reached 91.0 percent, but these require large annotated datasets and powerful hardware. The new pipeline, by contrast, ran on an ordinary AMD Ryzen 7 system with 16 gigabytes of RAM in MATLAB, without GPU acceleration. Segmentation took roughly four seconds per image due to its iterative nature, while feature extraction and classification across the entire 200-image dataset required only about four seconds in total, or roughly 0.02 seconds per image. The authors note that their method achieved better accuracy and specificity than some deep learning approaches, including Inception, SqueezeNet, and VGG-19, although customized deep networks still lead in raw performance.

The practical appeal of the work lies in its suitability for resource-constrained environments, where deep learning infrastructure is often unavailable. Because the pipeline needs no large-scale training and runs on commodity hardware, the researchers suggest it could serve as a valuable tool for primary screening of pneumonia and as a decision-support system assisting radiologists in differentiating bacterial from viral infections in pediatric patients. The team plans to extend the framework with ensemble and hybrid learning approaches and explainable techniques, and to validate it on larger and more diverse datasets to assess generalizability. For now, the study offers a concrete demonstration that two anatomical regions are better than one, and that carefully chosen shape descriptors can rival heavyweight artificial intelligence at a fraction of the computational cost.

Subject of Research: Fusion of lung and mediastinum imaging biomarkers for differentiating bacterial and viral pneumonia in pediatric chest radiographs

Article Title: Detection of bacterial and viral pneumonia in pediatric chest radiographs using fusion of mediastinum and lung imaging biomarkers

Article References: Tulo, S. K., Kerketta, S. R., Rahul Manohar, O., & Martha, P. (2026). Detection of bacterial and viral pneumonia in pediatric chest radiographs using fusion of mediastinum and lung imaging biomarkers. Frontiers in Signal Processing, 6, Article 1831207. https://doi.org/10.3389/frsip.2026.1831207

Image Credits: AI Generated

DOI: 10.3389/frsip.2026.1831207

Keywords: pneumonia, pediatric chest radiographs, bacterial pneumonia, viral pneumonia, mediastinum, image segmentation, level set methods, Hu moments, biomarker fusion, machine learning, support vector machine, linear discriminant analysis

The research, published in the journal Environmental Monitoring and Assessment, was led by Farzaneh Razmju of the University of Aberdeen together with colleagues in the United Kingdom, Hong Kong, India, Afghanistan and the United States. The team set out to answer a deceptively simple question: has tree cover on agricultural land in eastern Afghanistan changed over a period spanning one of the most dramatic political transitions in the country’s recent history? The study window, from 2016 to 2024, encompasses the collapse of the Afghan republican government in August 2021, the withdrawal of much international development assistance, and the restructuring of the institutions that had previously supported rural livelihoods and tree planting programmes.

To track change on the ground without setting foot in the field, the researchers turned to satellites. They used imagery from PlanetScope, a commercial constellation of small satellites that images nearly the entire land surface of the Earth at a resolution of 3 metres per pixel. That level of detail is fine enough to distinguish individual orchards, scattered farm trees and field boundaries, which is essential when the objects of interest are small patches of woody vegetation embedded in a mosaic of crops, settlements and river corridors. Scenes were selected from the mid-to-late growing season, in August and September, when orchard crops such as grapes and pomegranates have fully developed canopies and the spectral signatures of cropland, tree cover and other land classes are most clearly separated.

The analytical pipeline combined object-based image analysis with a machine learning classifier. Rather than classifying each pixel independently, which tends to produce speckled salt-and-pepper maps in high-resolution imagery, the team first segmented each image into homogeneous objects using the Segment Mean Shift algorithm, applying identical parameters across all four districts and all three study years. A support vector machine classifier, chosen for its robustness when training data are limited and landscapes are heterogeneous, then assigned each object to one of five classes: tree cover, cropland, water, built-up areas and barren land. The classifier’s two key hyperparameters, C and gamma, were optimised automatically through a two-dimensional grid search for each district-year combination, allowing the model to adapt to local spectral conditions.

Accuracy was assessed rigorously. For each classified map, 500 stratified random validation points were compared against reference data derived from visual interpretation of high-resolution historical imagery in Google Earth Pro, following established good practices in remote sensing. Every map achieved a Cohen’s kappa coefficient above 0.80, the threshold generally considered to indicate strong to almost perfect agreement with reference data. The highest value, 0.93, was recorded in Dehbala in 2020, while the lowest, 0.80, was recorded in Dara Noor in the same year. Post-classification change detection and transition matrices then quantified exactly which land classes had converted into which, and Sankey diagrams, in which flow widths are proportional to converted area, made the magnitude and direction of those transitions visually intuitive.

The headline result is unambiguous: tree cover increased in all four districts. The study focused on Dara Noor, Nazyan, Dehbala and Pachir Aw Agam, four predominantly mountainous districts whose cultivated inter-montane valleys collectively represent the irrigated, rain-fed and mixed cropland-orchard farming systems typical of eastern Afghanistan. In 2016, Dehbala held the largest tree cover extent at 216.96 hectares, or 10.21 percent of the analysed area, while Nazyan and Dara Noor started from very low baselines of around 3 to 4.6 percent. By 2024, tree cover had expanded everywhere, but the pace and pattern varied considerably. Nazyan recorded the most dramatic relative increase, with tree cover rising from 3.10 percent of the analysed area in 2016 to 17.03 percent in 2024, while Dara Noor showed a gradual but steady climb and Dehbala and Pachir Aw Agam displayed more moderate, fluctuating trends.

The transition analysis revealed where the new trees came from. Cropland was by far the most persistent land class throughout the study period, and built-up areas and water bodies also showed high stability. The dominant dynamic pathway was the direct conversion of cropland to tree cover, with smaller contributions from barren land and minimal transitions from built-up areas. Notably, the magnitude of cropland-to-tree conversion was more pronounced in the 2020 to 2024 sub-period than in the years before, indicating increased land cover dynamism in the later period, precisely the years following the political transition of 2021. This timing is intriguing, because it coincides with the withdrawal of many donor-supported agricultural development programmes, including the National Horticulture and Livestock Project, which had promoted orchard establishment, fruit tree sapling distribution and farmer training during the Republic period from 2012 to 2021.

What explains the greening of Afghan farmland amid national deforestation? The authors are careful to lay out several plausible mechanisms without claiming that any single one is confirmed. Deliberate tree planting by farmers, encouraged by the legacy of pre-2021 government and donor programmes promoting fruit trees, poplar and other multipurpose species, is one candidate. Natural regeneration of woody vegetation on land where cultivation has been reduced or temporarily abandoned is another; such processes are well documented in conflict-affected settings, meaning that cropland-to-tree transitions do not necessarily signal active agroforestry adoption. The satellite data alone cannot distinguish between newly established orchards and passive regrowth, and the relative contributions of these mechanisms remain an open question that would require field-based validation, which is difficult and dangerous in insecure environments.

The Nazyan district illustrates both the promise and the puzzle of the findings. Between 2016 and 2024, tree cover there rose from 3.10 to 17.03 percent of the analysed area, while the area classified as water increased from 9.75 to 21.29 percent, from 40.26 to 87.96 hectares, and cropland declined substantially from 50.88 to 32.77 percent. These concurrent changes point to a substantial reorganisation of land cover, but the study lacked the hydrological, precipitation and irrigation data needed to determine whether the expansion of mapped water was connected to the expansion of trees. Visual inspection of the high-resolution imagery shows both scattered trees within and along field boundaries and larger patches consistent with orchard systems, but the imagery cannot reliably separate deliberate planting from natural recovery.

Whatever the mechanism, the practical significance is considerable. Trees on farms provide fruit, fuelwood, fodder, shade and income, reduce soil erosion, enhance water retention and improve soil fertility, and can buffer remaining forest fragments while contributing to carbon storage. In a country where fuelwood scarcity remains a persistent challenge for rural households and where forests in the eastern provinces have suffered severe losses from illegal logging and conflict-related disturbance, farm-grown tree products can substitute for resources that would otherwise be extracted from natural forests. Studies from other regions, including Indonesia, have found that farmers practising agroforestry are less involved in forest clearing and forest product collection than those relying on swidden agriculture, lending weight to the hypothesis that Afghanistan’s on-farm tree expansion partly reflects greater reliance on farm-based resources, although the authors stress that this remains a hypothesis requiring household-level investigation.

The study also carries a methodological message for the world’s fragile states. Ground-based monitoring, reporting and verification of agroforestry investments are often impractical where security concerns and mountainous terrain block access, as demonstrated by evidence from Tigray in northern Ethiopia, where armed conflict was associated with vegetation degradation and disrupted restoration efforts. High-resolution satellite platforms such as PlanetScope, Sentinel-2 and Landsat offer consistent, repeatable and objective data even where researchers cannot go, and this study shows that they can be extended beyond natural forests to the trees outside forests that sustain rural livelihoods. The authors caution that their results describe medium-term change within four delineated agricultural areas and should not be generalised to all of Afghanistan, and that future work integrating field observations, household surveys and multi-season imagery will be needed to determine whether the observed gains represent a persistent trajectory or a shorter-term fluctuation. But as a demonstration that substantial tree cover change can occur within agricultural landscapes even amid continued landscape-level deforestation, the findings offer a rare note of measured optimism for post-conflict environmental recovery.

Subject of Research: Tree cover change on agricultural land in conflict-affected eastern Afghanistan monitored with high-resolution satellite imagery

Article Title: Tree cover increases on agricultural land amid continued landscape-level deforestation: a case study from eastern Afghanistan

Article References: Razmju, F., Gui, B., Sam, L., Bhardwaj, A., Haider, M. W., Khurram, S., Groninger, J. W., & Burslem, D. F. R. P. (2026). Tree cover increases on agricultural land amid continued landscape-level deforestation: a case study from eastern Afghanistan. Environmental Monitoring and Assessment, 198(11), Article 1203. https://doi.org/10.1007/s10661-026-16018-1

Image Credits: AI Generated

DOI: 10.1007/s10661-026-16018-1

Keywords: Afghanistan, tree cover, agroforestry, remote sensing, land use change, PlanetScope, Nangarhar, conflict, deforestation, satellite imagery, object-based image analysis, support vector machine

Soil aggregate stability, often abbreviated SAS, describes the ability of sand, silt and clay particles to remain bound into durable clumps when battered by raindrops, runoff or wind. It is far more than an abstract soil science metric. Stable aggregates create the interconnected pore networks that let water infiltrate, roots penetrate and microbes thrive, while unstable aggregates shed fine particles that clog pores, accelerate runoff and strip nutrients from the land. Scientists commonly quantify this property through the mean weight diameter, or MWD, obtained by sieving soil through a stack of mesh screens submerged in water and weighing what survives on each level.

The trouble is that measuring MWD in the laboratory is slow, labor-intensive and difficult to standardize. For a region like the eastern Himalayas, where plot-scale erosion data are scarce and terrain is treacherous, waiting on wet-sieving analyses for every field is simply impractical. That is where pedotransfer functions come in: statistical or machine learning models that estimate hard-to-measure soil properties from easily obtained ones such as texture, organic carbon, bulk density and pH. The new study, published in Frontiers in Soil Science, is the first to build such functions specifically for aggregate stability in this fragile mountain region.

Led by researchers from the Indian Council of Agricultural Research and partner institutions, the team collected 126 topsoil samples from the top 15 centimeters of a 113-square-kilometer sub-watershed called Ranikhola, spanning agricultural fields and forest land between elevations of 1,200 and 1,500 meters. Sixty-one samples came from farmland and sixty-five from forest. In the laboratory, the researchers measured ten physico-chemical properties for each sample: the percentages of sand, silt and clay, organic carbon content, bulk density, pH, and the concentrations of calcium, magnesium, carbonate and bicarbonate ions. These ten variables served as inputs for the models, with MWD as the prediction target.

The soil chemistry told a revealing story before any modeling began. Agricultural soils were slightly denser, with a mean bulk density of 1.25 grams per cubic centimeter against 1.17 for forest soils, and contained less organic carbon, 3.37 percent versus 4.23 percent. Farmland was sandier, averaging 61.34 percent sand, while forest soils held more clay, 30.11 percent. Both land uses fell into the medium stability class, with mean MWD values of 1.09 and 1.11 millimeters respectively, but the forest edge was consistent with its richer organic matter, which acts as a natural glue binding particles into aggregates. Statistical tests confirmed that pH, combined calcium and magnesium, organic carbon and bulk density differed significantly between the two land uses.

Correlation analysis identified clay content, organic carbon and bulk density as the dominant drivers of aggregate stability, echoing decades of soil science showing that clay particles and organic matter function as the principal binding agents. Intriguingly, the analysis also revealed a negative relationship between pH and MWD, consistent with the idea that in acidic soils below pH 7, heightened microbial activity promotes macro-aggregation while cations become more mobile and available to bridge clay and organic matter.

With the dataset assembled, the team pitted four modeling approaches against one another: classical multiple linear regression, or MLR, alongside three machine learning heavyweights — artificial neural networks, random forests and support vector machines. The dataset was split 80:20 into training and testing sets, with ten-fold cross-validation applied during model development. To guard against the luck of any single data split, the entire procedure was repeated fifty times with different random partitions, and performance metrics were averaged across all repetitions. Models were judged on the coefficient of determination, root mean square error, mean absolute error, Nash-Sutcliffe efficiency and the Akaike Information Criterion, which balances fit against complexity.

The results were striking. Multiple linear regression, which assumes straight-line relationships between variables, fell short across the board, with testing R² values as low as 0.13 for agricultural soils — well below the threshold for satisfactory prediction. The machine learning models, by contrast, captured the tangled non-linear interactions that govern how soil particles bind. Random Forest, an ensemble of decision trees that averages hundreds of individual predictions to suppress overfitting, delivered the best performance on agricultural land, achieving a training R² of 0.89 and a testing R² of 0.66. The neural network, a multi-layer perceptron whose hidden layers learn abstract representations of the input data, reigned supreme on forest soils, posting a remarkable R² of 0.92 in training and 0.91 in testing — nearly unchanged between the two phases, a sign of genuine generalization rather than memorization.

But the most intriguing result came from the support vector machine, which maps input data into a high-dimensional feature space where non-linear boundaries become linear, guided by a radial basis function kernel. When trained on the pooled dataset combining both land uses, SVM achieved a training R² of 0.98, the highest of any model in the study, with a testing R² of 0.72. While no single algorithm dominated every scenario, SVM was either the best or second-best performer in all cases, suggesting it possesses an unusual capacity to generalize across heterogeneous soil conditions. The researchers attribute this to the structural risk minimization principle underlying SVM, which balances fitting the training data against maintaining predictive power on unseen samples.

The practical implications reach well beyond the laboratory. India’s national Soil Health Card program already generates vast repositories of basic soil measurements from farms across the country. A validated pedotransfer function like the one developed here could convert those routine records into erosion susceptibility estimates without a single additional wet-sieving test, feeding directly into regional erosion models and digital soil erosion potential maps for data-scarce stretches of eastern and northeastern India. For a state where water erosion affects more than 38 percent of the land area, with degradation ranging from under 5 to over 80 tonnes per hectare per year, that shortcut could be decisive for targeting conservation resources where they matter most.

The authors are candid about the study’s limits. The 126 samples came from a single sub-watershed during one winter season, and only ten physico-chemical variables were considered — biological, mineralogical, topographic and management factors that also shape aggregation were left out. External validation on independent Himalayan datasets remains to be done. Future work, they suggest, should span multiple seasons and soil depths, incorporate terrain attributes and remote sensing products, and explore explainable AI and deep learning architectures with formal uncertainty quantification. Coupling the aggregate stability models with watershed-scale erosion and sediment transport simulators would complete the pipeline from prediction to prevention.

Still, the core message is resonant: in one of the world’s most erosion-vulnerable landscapes, the intelligence needed to protect the soil may already be sitting in routine soil test reports, waiting for the right algorithm to unlock it. As machine learning continues its march through the geosciences, the humble soil crumb — and the algorithms that now read it — may become unlikely heroes of mountain sustainability.

Subject of Research: Machine learning prediction of soil aggregate stability for erosion assessment in the eastern Himalayas

Article Title: Machine learning-based prediction of soil aggregate stability for sustainable land management in the eastern Himalayan region of India

Article References: Gupta, S., Yadav, P., Panda, J., Behura, K. B., Ghosh, S., Singh, L. K., Devi, S. R., Ghosh, M., Mohanty, A. K., & Pathak, P. K. (2026). Machine learning-based prediction of soil aggregate stability for sustainable land management in the eastern Himalayan region of India. Frontiers in Soil Science, 6, Article 1937878. https://doi.org/10.3389/fsoil.2026.1937878

Image Credits: AI Generated

DOI: 10.3389/fsoil.2026.1937878

Keywords: soil aggregate stability, machine learning, pedotransfer function, mean weight diameter, soil erosion, eastern Himalayas, Sikkim, random forest, support vector machine, artificial neural network, soil organic carbon, sustainable land management

The choice of a support vector machine for the feature-level fusion stage deserves some contextualization. Support vector machines construct decision boundaries by maximizing the margin between class clusters in a transformed feature space, and they remain competitive with deep neural networks when the number of training samples is small relative to the feature dimensionality. In industrial condition-monitoring experiments conducted on physical test rigs, dataset sizes are almost always constrained by the cost of running controlled fault scenarios, since each fault condition must be physically induced or allowed to develop under supervision. Under such constraints, kernel-based classifiers with well-engineered features frequently outperform end-to-end learned models, which require far more data to generalize. The 98.27 percent macro-F1 score achieved by the fused feature representation with an SVM classifier is consistent with this pattern, and the fact that decision-level fusion pushed performance above 99.6 percent indicates that the two modalities carry largely non-redundant information that can be combined either at the feature vector stage or at the classifier output stage with comparable success.

Macro-F1 as an evaluation metric warrants brief explanation for readers accustomed to overall accuracy figures. Macro-averaging computes the F1 score independently for each class and then takes the unweighted mean, which means that a classifier cannot achieve a high macro-F1 by performing well on frequent classes while neglecting rare ones. In fault diagnosis, where the six classes here comprise one normal state and five distinct component faults, macro-F1 is a more demanding criterion than accuracy because it penalizes poor performance on any individual fault category equally, regardless of how many samples that category contributes. The reported improvements of 7.06 percentage points over the current-only model and 14.50 percentage points over the vibration-only model therefore represent gains that are distributed across all fault categories rather than concentrated in the easiest ones, which strengthens the claim that fusion provides broadly useful discriminative information.

The fivefold contiguous time-block cross-validation protocol is a methodologically important detail that distinguishes careful condition-monitoring studies from those that overstate their results. In many published machine-learning diagnostics, samples from a single recording session are randomly shuffled and split into training and test folds, which allows the model to encounter nearly identical signal segments in both sets and produces inflated performance estimates. Contiguous time-block validation instead partitions the recording timeline into consecutive blocks and holds out entire blocks for testing, ensuring that the model must classify signal segments that are temporally separated from anything it saw during training. This design tests whether the learned feature representations capture stable physical signatures of each machine state rather than transient artifacts of a particular moment. The use of this protocol here, while still confined to a single experimental batch, means the reported scores reflect genuine temporal generalization within the batch, which is a more meaningful benchmark than random-split results would be.

The feature-group ablation findings align with established understanding of vibration-based diagnostics. Time-domain features, which include statistical moments, peak values, and crest factors computed directly from the raw acceleration signal, are sensitive to the overall energy and impulsiveness of the vibration and respond quickly to gross changes in machine behavior. Frequency-domain features, typically derived through spectral estimation methods such as Welch’s averaged periodogram, resolve the signal into its constituent frequency components and can isolate characteristic frequencies associated with specific rotating elements, pump pulsation harmonics, or structural resonances. Cross-axis information captures the directional structure of the vibration, since different fault mechanisms excite different geometric directions: a bearing defect may manifest predominantly in the radial plane, while an axial misalignment or thrust-related anomaly produces stronger axial response. The ablation results confirming that all three feature groups contribute to recognition performance suggest that the fault conditions studied produce distinguishable signatures across multiple descriptive dimensions, and that no single conventional feature family is sufficient on its own.

The physical rationale for combining vibration with supply current rests on the relationship between electrical power consumption and mechanical load. In an induction-motor-driven high-pressure pump, the current drawn from the supply reflects the torque demanded by the pump, which in turn depends on the pressure being generated and the volumetric flow being delivered. When abrasive flow rate changes, when the nozzle orifice wears and alters the discharge coefficient, or when internal pump leakage develops, the mechanical load shifts and the current signature follows. Vibration, by contrast, responds to the dynamic forces and impacts within each component, including pressure pulsations, particle impacts, mechanical looseness, and rubbing or bearing distress. Because these two quantities respond to overlapping but distinct sets of physical mechanisms, their combination provides a more complete description of machine state than either alone. The current modality acts as a slowly varying envelope that contextualizes the faster vibration fluctuations, helping the classifier distinguish, for example, a vibration pattern caused by a genuine component fault from a similar pattern caused by a temporary change in operating load.

The four monitoring locations were selected to cover the principal functional units of the machine. The high-pressure pump is the heart of the system, converting mechanical power into water pressure through an intensifier mechanism, and its dynamic behavior dominates the pressure waveform delivered downstream. The nozzle end sits closest to the cutting action, where jet formation, abrasive entrainment, and mixing occur, making it sensitive to nozzle wear, abrasive feed irregularities, and cutting-head blockage. The machine-tool spindle carries the motion system’s structural dynamics and reflects positioning accuracy and mechanical integrity of the drive train. The abrasive-tank bottom pressure-port region monitors the abrasive delivery system, where flow disruption or tank pressurization anomalies originate. Placing triaxial accelerometers at these four points yields twelve vibration channels that span the machine’s process chain, allowing the classifier to learn location-specific signatures and, through the rule-based module, to verify that predicted fault states are physically consistent with the monitoring locations where the strongest deviations are observed.

The deterministic rule-based post-classification module addresses a persistent gap between academic diagnostic models and practical industrial deployment. In real maintenance workflows, a bare classification output, however accurate, provides limited actionable guidance: maintenance personnel need to know what evidence supports the diagnosis, which other locations should be checked to confirm or refute it, and what verification steps to take. By applying a fixed topology matrix that encodes the physical relationships between components and monitoring points, together with predefined evidence rules that organize observed feature deviations into structured records, the module converts a probabilistic classifier output into an auditable diagnostic report. Crucially, the module preserves the original classifier outputs rather than modifying them, which maintains a clean separation between statistical inference and rule-based reasoning. This separation means that errors in the rule layer cannot corrupt the classification itself, and that the rule layer can be inspected, audited, and updated independently of the trained model, which is valuable in regulated or safety-critical industrial environments.

The explicit acknowledgment that the reported performance represents within-batch recognition under fixed sensor placement is an important caveat that deserves elaboration. Sensor remounting, even with magnetic attachment at the same nominal location, introduces variability in coupling stiffness, orientation, and contact geometry that can shift the vibration transfer function between the source and the sensor. Different machines of the same model exhibit manufacturing tolerances, assembly variations, and wear histories that alter their dynamic signatures. Operating batches may differ in material type, cutting parameters, abrasive characteristics, and ambient conditions, all of which modulate the signals available to the classifier. A model trained on one batch may therefore lose accuracy when applied to another, a phenomenon related to domain shift in machine learning. The authors’ careful scoping of their claims to within-batch recognition, rather than claiming cross-machine or cross-batch generalization, reflects a rigorous understanding of these limitations and provides an honest baseline against which future transfer-learning and domain-adaptation studies can be measured.

The abrasive waterjet machining context itself shapes the diagnostic challenge in specific ways. Unlike conventional machining, where tool wear progresses gradually and cutting forces provide a direct measure of tool-workpiece interaction, AWJ machining involves a high-velocity mixture of water and abrasive particles whose cutting action depends on jet coherence, particle velocity distribution, and standoff distance. The absence of a rigid tool contact means that force-based monitoring is less straightforward, and the high-pressure environment limits sensor placement options near the cutting zone. Vibration sensing on the machine structure offers a non-intrusive alternative that captures the structural response to jet formation and component operation without exposing instruments to the abrasive stream. Combined with electrical current measurement, which requires only a clamp meter on the supply circuit and involves no contact with the process at all, the two modalities form a practical sensing pair that can be retrofitted to existing machines with minimal modification.

The sampling configuration also merits attention. Vibration was acquired at 1 kHz per channel, a rate sufficient to capture frequency content up to the Nyquist limit of 500 Hz, which encompasses the dominant structural resonances and pump pulsation frequencies of typical high-pressure systems. The current signal was sampled at a median interval of 0.75 seconds, a much slower rate appropriate for a modality that varies slowly with load changes rather than with mechanical oscillations. Aligning these heterogeneous sampling rates into common 2-second analysis windows required careful synchronization, and the resulting dataset pairs each vibration snapshot with the corresponding current reading, enabling the fusion models to learn joint patterns across timescales. This multi-rate acquisition strategy mirrors practical industrial deployments, where vibration monitoring systems and power monitoring systems typically operate independently with different sampling characteristics, and demonstrates that meaningful fusion is achievable without forcing all modalities onto a common high-rate clock.

Looking toward future development, the framework establishes several natural directions for extension. Cross-batch validation, in which the model is trained on one experimental campaign and tested on a separate campaign conducted under different conditions, would directly probe the generalization limits that the current study deliberately excludes. Transfer-learning techniques, including domain adaptation and fine-tuning strategies, could address the variability introduced by sensor remounting and machine-to-machine differences. Expanding the fault taxonomy beyond the five predefined conditions to include progressive degradation states rather than binary fault categories would enable prognostic capabilities alongside diagnostic ones. The structured evidence records produced by the rule-based module could additionally serve as training material for higher-level reasoning systems, bridging the gap between statistical classification and the maintenance decision-making processes used in industrial practice, while the deterministic architecture ensures that such integration does not compromise the transparency and auditability of the diagnostic chain.

Subject of Research: Multimodal vibration and current fusion for recognizing component faults in an abrasive waterjet machine tool

Article Title: Vibration–current multimodal fusion for within-batch recognition of predefined component faults of a high-pressure abrasive waterjet machine tool

Article References: Zhang, G., Zhang, C., Jiang, H., Zhang, X., Wang, Y., & Bao, X. (2026). Vibration–current multimodal fusion for within-batch recognition of predefined component faults of a high-pressure abrasive waterjet machine tool. Frontiers in Mechanical Engineering, 12, Article 1970496. https://doi.org/10.3389/fmech.2026.1970496

Image Credits: AI Generated

DOI: 10.3389/fmech.2026.1970496

Keywords: abrasive waterjet machining, fault diagnosis, multimodal fusion, vibration monitoring, machine current, support vector machine, cross-validation, condition monitoring, feature ablation, rule-based evidence, Vibration, current

The stakes are considerable. Cities are expanding at unprecedented pace as the global population climbs from about 8 billion toward a projected 10.4 billion by the end of the century, and planners increasingly rely on satellite-derived land-cover maps to track impervious surfaces, urban heat islands, vanishing green space, and flood-prone peri-urban fringes. Yet the sensors that supply this information see the world very differently, and the new research demonstrates that those differences ripple all the way through to the maps on which planning decisions rest.

The team, led by Sawaid Abbas of the University of the Punjab and The Hong Kong Polytechnic University, classified Landsat-8, Sentinel-2, and PRISMA imagery within a harmonized framework: identical reference samples, the same eight classes, and a common validation protocol. The classes spanned an urban intensity gradient from Dense and Medium to Sparse Urban Area, alongside Tree Cover, Grass, Cropland, Bare Area, and Water. Five machine-learning classifiers were tested on each sensor: Random Forest, Support Vector Machine, Gradient Boosting Machine, K-Nearest Neighbors, and Classification and Regression Trees, with performance judged by overall accuracy, Cohen’s kappa, and F1-scores.

The technical gap between the sensors is stark. Landsat-8 offered seven broad spectral bands at 30 meters; Sentinel-2 contributed ten bands, including red-edge wavelengths sensitive to vegetation, at 10 to 20 meters; PRISMA captured 239 narrow hyperspectral bands spanning 400 to 2500 nanometers at 30 meters, of which 187 remained after noisy water-vapor-affected bands were removed and a Savitzky-Golay filter cut median spectral roughness by nearly 96 percent. For PRISMA, principal component analysis compressed the 187 bands into 15 components retaining 99.9 percent of the training variance, taming the curse of dimensionality that plagues hyperspectral classification.

The results were dramatic. PRISMA paired with a tuned radial-basis-function Support Vector Machine achieved 91.26 percent overall accuracy and a macro-averaged F1-score of 89.72 percent, beating the best Landsat result by 5.21 percentage points and the best Sentinel-2 result by 8.12 points. The single largest gain came in the transitional Medium Urban class, whose F1-score leapt from roughly 47 percent on both multispectral sensors to 86.15 percent with PRISMA, an improvement of nearly 40 points in the very category that matters most for monitoring informal expansion and peri-urban sprawl.

Spectral separability analysis explained why. Medium and Sparse Urban surfaces are nearly indistinguishable in Landsat data, separated by a spectral angle of just 0.54 degrees, and only marginally better in Sentinel-2 at 1.58 degrees. PRISMA widened that angle to 4.11 degrees and boosted the standardized centroid distance to 6.13, providing hard quantitative evidence that dense spectral sampling resolves mixed pixels containing rooftops, bare soil, and vegetation that broad-band sensors blur together. Interestingly, the most informative PRISMA bands clustered in the near-infrared range between 809 and 830 nanometers.

Yet the study refuses to crown a universal winner. PRISMA’s classifications redistributed large areas from Sparse Urban into Bare Area and Cropland, which the authors interpret as plausibly better discrimination of mixed peri-urban surfaces, but they caution that this cannot be declared more accurate without an independent area-wide reference map. A six-month acquisition gap between the February and March multispectral scenes and the August PRISMA image introduces seasonal differences in vegetation and soil moisture, and PRISMA classifiers were extensively tuned while the multispectral runs used standard cloud-platform settings, both factors that could shape the apparent advantage.

The comparison also highlights complementary strengths. Landsat performed best for dense built-up surfaces, tree cover, grass, and bare soil, and remains irreplaceable for its five-decade archive of urban change. Sentinel-2, though it scored lowest here at 83.14 percent overall accuracy under Random Forest, delivered the finest spatial rendering of fragmented urban morphology, and its red-edge bands proved highly sensitive to vegetation variation, with grass estimates swinging from 164 to 280 square kilometers depending on the classifier.

Framed as a foundation for what the authors call urban digital intelligence, the study argues that resilient city planning needs all three sensing dimensions at once: Landsat’s temporal continuity, Sentinel-2’s spatial detail, and the spectral richness of hyperspectral missions such as PRISMA, EnMAP, ESA’s forthcoming CHIME, and NASA’s Surface Biology and Geology program. Better maps of transitional urban-density zones could sharpen heat-mitigation corridor design, green-infrastructure investment, and monitoring toward the UN’s sustainable development target for cities.

The authors acknowledge open questions, including the need for spatial block validation to rule out autocorrelation between training and validation samples, and the promise of deep-learning architectures that could fuse multi-sensor data more completely. But the core message stands: urban land-cover intelligence is inherently multidimensional, and as hyperspectral missions multiply, the era of relying on any single satellite to understand a city is drawing to a close.

Subject of Research: Cross-sensor comparison of multispectral and hyperspectral satellite imagery for urban land-cover classification

Article Title: Cross-sensor urban land-cover characterization using hyperspectral and multispectral satellite imagery

Article References: Abbas, S., Zohaib, Habib, N., Qamer, F. M., & Nazeer, M. (2026). Cross-sensor urban land-cover characterization using hyperspectral and multispectral satellite imagery. Frontiers in Remote Sensing, 7, Article 1908768. https://doi.org/10.3389/frsen.2026.1908768

Image Credits: AI Generated

DOI: 10.3389/frsen.2026.1908768

Keywords: remote sensing, hyperspectral imaging, PRISMA, Landsat-8, Sentinel-2, urban land cover, machine learning, land-cover classification, Support Vector Machine, urban planning, spectral separability, Random Forest

The research, led by Zhongyuan Yan and colleagues at Henan Jiuyu Longyuan Energy Service Co., Ltd. in Zhengzhou, China, addresses a persistent blind spot in power system planning. In most existing studies, renewable power forecasting and storage planning are treated as separate problems. Forecasting researchers focus on shaving prediction errors, while storage planners simply plug in predetermined renewable power profiles as fixed inputs. The consequence is that forecasting errors propagate silently into storage sizing decisions, potentially leading to systems that are either too small to absorb surplus renewable generation or too large to justify their capital cost.

The team’s solution begins at the data level with a forecasting model they call VMD-EO-SVM, a hybrid that chains together three distinct techniques. Variational mode decomposition, or VMD, first breaks down the raw wind or solar power time series into a set of narrow-band intrinsic mode functions, each capturing fluctuations at a different frequency scale. This decomposition tames the non-stationary, erratic character of renewable output by separating fast spikes from slow trends, effectively turning one hard prediction problem into several easier ones.

Next comes the equilibrium optimizer, a nature-inspired metaheuristic that mimics the physics of mass exchange in a control-volume system seeking equilibrium. The EO searches for the optimal values of the penalty parameter C and the kernel width gamma in a support vector machine, the third component of the pipeline. SVM regression maps input samples into a high-dimensional feature space where a linear regression function is constructed, using a radial basis function kernel prized for its wide convergence region and high resolution. By letting EO tune these hyperparameters for each decomposed component, the framework avoids the guesswork that typically undermines SVM performance. The component-level forecasts are then reconstructed into a single final prediction.

The benchmark results are striking. Using 15-minute-interval measured wind and solar power data from Belgium covering 2020 to 2021, the VMD-EO-SVM model achieved a root mean square error of 42.60 for solar power with a coefficient of determination of 0.9776, and an RMSE of 71.83 for wind power with an R-squared of 0.9217. Compared against deep learning benchmarks, the hybrid model cut solar RMSE by 38.79 percent relative to LSTM and 55.09 percent relative to CNN, and reduced wind RMSE by 37.95 and 50.68 percent against the same rivals. Even more telling is the component-level ablation: for solar power, RMSE fell from 195.21 for plain SVM to 56.08 for EO-SVM and then to 42.60 for the full VMD-EO-SVM stack, demonstrating that decomposition and parameter optimization perform complementary functions rather than redundant ones.

Notably, the study found that signal decomposition alone does not guarantee improvement. For wind power, VMD-SVM actually failed to outperform the original SVM in RMSE terms. It is only when VMD is paired with EO-optimized SVM that the combination delivers its full benefit, suggesting that the two techniques must work in concert: VMD reduces the complexity of the prediction task while EO ensures the SVM parameters are properly matched to each frequency-separated component.

With reliable forecasts in hand, the second stage of the framework turns to the planning question: how big should a shared energy storage facility be? Unlike storage systems configured independently for individual distributed energy systems, shared energy storage allows multiple generation and consumption entities to draw on a common facility. Because different subsystems have partially complementary charging and discharging patterns, aggregation reduces redundant capacity and improves utilization. The capital burden is also distributed among participants rather than borne by any single user, a mechanism the authors argue is essential for making storage investment economically viable at cluster scale.

The multi-objective planning model minimizes total system cost, which includes the daily-equivalent investment cost of storage, operation and maintenance expenses, thermal generation fuel costs, and penalties for renewable curtailment, while simultaneously maximizing the renewable energy utilization ratio. Constraints enforce power balance at every time step, limit curtailment to predicted renewable output, keep the storage state of charge between 0.1 and 0.9, and respect the ramp-rate and minimum on-off-time requirements of seven thermal units ranging from 55 to 130 MW. The optimization determines both the rated power and the energy capacity of the shared storage system.

The case study results reveal a configuration of 947 MW rated power and 5,615 MWh of energy capacity, corresponding to an energy-to-power ratio of approximately 5.93 hours. That duration is significant: it indicates the storage system is not merely smoothing second-to-second fluctuations but shifting energy across several hours, absorbing midday solar surpluses and overnight wind output for release during peak demand. The optimized plan achieves a renewable energy accommodation rate of 90.3 percent with a total system cost of 66,956,684 CNY, of which 64.3 million represents the daily-equivalent storage investment. Operational simulations confirm that the storage charges during periods of renewable surplus and discharges when generation falls short or load peaks, with the state of charge never straying outside safe bounds.

The thermal fleet plays a supporting role in this choreography. Units 1 and 2 operate stably as base-load generators, while units 3 through 7 act as peaking units that shut down during high renewable generation to cut coal consumption, keeping startup and shutdown costs to just 41,720 CNY. Wind and solar output show clear temporal complementarity, with solar dominating daytime hours and wind carrying the night, allowing the combined system to maintain real-time power balance across all 96 fifteen-minute periods of the scheduling day.

The authors are candid about limitations. The forecasting model relies solely on historical power data and does not explicitly incorporate meteorological variables such as wind speed, solar irradiance, or temperature. While historical series implicitly encode past weather patterns, this indirect representation may prove inadequate when future conditions diverge sharply from historical norms, particularly during extreme weather events. Future work, they note, will integrate numerical weather prediction directly into the forecasting framework and examine how extreme conditions ripple through to storage planning decisions.

Beyond the technical achievements, the study carries a broader message for the energy transition: forecasting and storage planning should be treated as a single coupled problem, not sequential silos. Better forecasts yield more reliable storage sizing, which in turn unlocks higher renewable penetration at lower cost. As wind-solar-thermal clusters proliferate under carbon peaking and neutrality targets worldwide, frameworks that bind prediction quality directly to infrastructure investment may well become the template for how distributed grids of the future are designed.

Subject of Research: Optimal planning of shared energy storage for wind-solar-thermal distributed energy system clusters using improved renewable power forecasting

Article Title: Optimal planning of shared energy storage for wind-solar-thermal distributed energy system clusters based on improved renewable energy forecasting

Article References: Yan, Z., Xue, W., Zhang, Y., Du, X., & Zheng, X. (2026). Optimal planning of shared energy storage for wind-solar-thermal distributed energy system clusters based on improved renewable energy forecasting. Frontiers in Energy Research, 14, Article 1946325. https://doi.org/10.3389/fenrg.2026.1946325

Image Credits: AI Generated

DOI: 10.3389/fenrg.2026.1946325

Keywords: shared energy storage, renewable energy forecasting, wind power, solar power, variational mode decomposition, equilibrium optimizer, support vector machine, distributed energy systems, multi-objective optimization, renewable energy accommodation, energy storage planning, smart grids