Leave-one-out Cross-validation Research Articles

Interpretable machine learning assisted multi-objective optimization design for small molecule hole transport materials

Small molecule hole transport materials (SM-HTMs) play an important role in the development of perovskite solar cells (PSCs) due to the simple structure and excellent photoelectric performance. Herein, we established interpretable models for designing materials with high hole mobility (μ) and absolute hardness (η). The μ and η models were constructed by Gradient Boosting Regression (GBR) due to the better performances. The coefficient of determination (R2) for leave-one-out cross validation (LOOCV) and the test set are 0.86 and 0.86 in μ model, 0.84 and 0.89 in η model, respectively. SHapley Additive exPlanations (SHAP) analysis reveals that higher μ are more likely to be obtained with fewer C–O and N-O pairs at topological distance 9, and molecules with high degree of cyclization and less polar fragments might have higher η. New groups were designed based on the strategies and used to generate 785 virtual samples. Finally, 35 promising candidates (μ > 10−1.54 cm2 V−1 s−1 and η > 2.71 eV) with high synthesizability were screened out using high-throughput screening and pattern recognition. Our work provides an effective way to accelerate the design of prominent SM-HTMs.

Enhanced Kriging leave-one-out cross-validation in improving model estimation and optimization

Leave-one-out cross-validation (LOOCV) is a widely used technique in model estimation and selection of the Kriging surrogate model for engineering problems, such as structural optimization and reliability analysis. However, the traditional LOOCV method has some disadvantages in terms of accuracy and efficiency. This paper proposes an enhanced-LOOCV method that incorporates hyperparameters from the Kriging model based on the complete training dataset (i.e., the complete Kriging model) into the LOOCV error calculation. By keeping the model hyperparameters in LOOCV consistent with the complete Kriging model, it reduces the number of hyperparameter optimizations and significantly increases the accuracy and efficiency of the LOOCV process. Additionally, a decremental calculation is proposed to reduce the computational costs of the correlation matrix inversion without sacrificing accuracy, resulting in an improved time complexity of the traditional LOOCV from O(n4) to O(n3). Experimental results with thirty test functions verify that the enhanced-LOOCV has better estimation performance than the Kriging model with significantly higher efficiency compared to the traditional LOOCV. Numerical experiments and an engineering case in optimization demonstrate that the enhanced-LOOCV can reduce the number of samples needed for infilling in the Kriging model, and it is more suitable for expensive optimizations in engineering.

Enhanced Differentiation of Wild and Feeding Civet Coffee Using Near-Infrared Spectroscopy with Various Sample Pretreatments and Chemometric Approaches

Civet coffee is the world’s most expensive and rarest coffee bean. Indonesia was the first country to be identified as the origin of civet coffee. First, it is produced spontaneously by collecting civet feces from coffee plantations near the forest. Due to limited stock, farmers began cultivating civets to obtain safe supplies of civet coffee. Based on this, civet coffee can be divided into two types: wild and fed. A combination of spectroscopy and chemometrics can be used to evaluate authenticity with high speed and precision. In this study, seven samples from different regions were analyzed using NIR Spectroscopy with various preparations: unroasted, roasted, unground, and ground. The spectroscopic data were combined with unsupervised exploratory methods (hierarchical cluster analysis (HCA) and principal component analysis (PCA)) and supervised classification methods (support vector machine (SVM) and random forest (RF)). The HCA results showed a trend between roasted and unroasted beans; meanwhile, the PCA showed a trend based on coffee bean regions. Combining the SVM with leave-one-out-cross-validation (LOOCV) successfully differentiated 57.14% in all sample groups (unground, ground, unroasted, unroasted–unground, and roasted–unground), 78.57% in roasted, 92.86% in roasted–ground, and 100% in unroasted–ground. However, using the Boruta filter, the accuracy increased to 89.29% for all samples, to 85.71% for unground and unroasted–unground, and 100% for roasted, unroasted–ground, and roasted–ground. Ultimately, RF successfully differentiated 100% of all grouped samples. In general, roasting and grinding the samples before analysis improved the accuracy of differentiating between wild and feeding civet coffee using NIR Spectroscopy.

Multivariate regression modelling for gender prediction using volatile organic compounds from hand odor profiles via HS-SPME-GC-MS.

The efficacy of using human volatile organic compounds (VOCs) as a form of forensic evidence has been well demonstrated with canines for crime scene response, suspect identification, and location checking. Although the use of human scent evidence in the field is well established, the laboratory evaluation of human VOC profiles has been limited. This study used Headspace-Solid Phase Microextraction-Gas Chromatography-Mass Spectrometry (HS-SPME-GC-MS) to analyze human hand odor samples collected from 60 individuals (30 Females and 30 Males). The human volatiles collected from the palm surfaces of each subject were interpreted for classification and prediction of gender. The volatile organic compound (VOC) signatures from subjects' hand odor profiles were evaluated with supervised dimensional reduction techniques: Partial Least Squares-Discriminant Analysis (PLS-DA), Orthogonal-Projections to Latent Structures Discriminant Analysis (OPLS-DA), and Linear Discriminant Analysis (LDA). The PLS-DA 2D model demonstrated clustering amongst male and female subjects. The addition of a third component to the PLS-DA model revealed clustering and minimal separation of male and female subjects in the 3D PLS-DA model. The OPLS-DA model displayed discrimination and clustering amongst gender groups with leave one out cross validation (LOOCV) and 95% confidence regions surrounding clustered groups without overlap. The LDA had a 96.67% accuracy rate for female and male subjects. The culminating knowledge establishes a working model for the prediction of donor class characteristics using human scent hand odor profiles.

A new approach for quantitative precipitation estimation from radar reflectivity using a gated recurrent unit network

Accurate quantitative precipitation estimation (QPE) utilizing radar data plays a vital role in weather monitoring, hydrological modelling and prediction. In previous research, some regression techniques such as the reflectivity (Z) - rainfall (R) relationship and artificial intelligence methods were employed for radar QPE; however, the estimation results remain limited the spatial and temporal variability of rainfall results in large errors in radar QPE. Recently, an emerging deep learning approach that can be applied in hydrological areas for QPE, namely, the gated recurrent unit (GRU) network method, has been used. GRU networks have evolved from long short-term memory (LSTM) networks, with improved training speed and modelling accuracy. In this research, the GRU network is applied to QPE using the three-dimensional structure of reflectivity from the Constant Altitude Plan Position Indicator (3D CAPPI) for an urban catchment in Seoul, South Korea. Simultaneously, a “complete” 3D CAPPI based on a cubic grid of 4 × 4 km2 that is highly suitable for the study area is proposed. Based on the comparison of the results of the GRU model with those of the artificial neural networks (ANNs), recurrent neural networks (RNNs), and LSTM networks, and the Z-R relationship, namely the least-squares model (LSM) for 2 testing strategies including traditional test strategy and leave-one-out cross-validation (LOOCV) strategy, the results indicate that the GRU model is superior to other models in radar QPE, with improvements by 3.3 ∼ 46.2%, 25 ∼ 70% and 1.2 ∼ 44.8% in the root mean square error (RMSE), bias and correlation coefficient (CC). This study presents a valuable method for improving the accuracy of radar QPE that can be actively employed in the future.

Quantitative structure-activity relationship study of amide derivatives as xanthine oxidase inhibitors using machine learning.

The target of the study is to predict the inhibitory effect of amide derivatives on xanthine oxidase (XO) by building several models, which are based on the theory of the quantitative structure-activity relationship (QSAR). The heuristic method (HM) was used to linearly select descriptors and build a linear model. XGBoost was used to non-linearly select descriptors, and radial basis kernel function support vector regression (RBF SVR), polynomial kernel function SVR (poly SVR), linear kernel function SVR (linear SVR), mix-kernel function SVR (MIX SVR), and random forest (RF) were adopted to establish non-linear models, in which the MIX-SVR method gives the best result. The kernel function of MIX SVR has strong abilities of learning and generalization of established models simultaneously, which is because it is a combination of the linear kernel function, the radial basis kernel function, and the polynomial kernel function. In order to test the robustness of the models, leave-one-out cross validation (LOOCV) was adopted. In a training set, = 0.97 and RMSE = 0.01; in a test set, = 0.95, RMSE = 0.01, and = 0.96. This result is in line with the experimental expectations, which indicate that the MIX-SVR modeling approach has good applications in the study of amide derivatives.

Prediction of early improvement of major depressive disorder to antidepressant medication in adolescents with radiomics analysis after ComBat harmonization based on multiscale structural MRI

BackgroundDue to individual differences and lack of objective biomarkers, only 30-40% patients with major depressive disorder (MDD) achieve remission after initial antidepressant medication (ADM). We aimed to employ radiomics analysis after ComBat harmonization to predict early improvement to ADM in adolescents with MDD by using brain multiscale structural MRI (sMRI) and identify the radiomics features with high prediction power for selection of selective serotonin reuptake inhibitors (SSRIs) and serotonin norepinephrine reuptake inhibitors (SNRIs).Methods121 MDD patients were recruited for brain sMRI, including three-dimensional T1 weighted imaging (3D-T1WI)and diffusion tensor imaging (DTI). After receiving SSRIs or SNRIs for 2 weeks, the subjects were divided into ADM improvers (SSRIs improvers and SNRIs improvers) and non-improvers according to reduction rate of the Hamilton Depression Rating Scale, 17 item (HAM-D17) score. Then, sMRI data were preprocessed, and conventional imaging indicators and radiomics features of gray matter (GM) based on surface-based morphology (SBM) and voxel-based morphology (VBM) and diffusion properties of white matter (WM) were extracted and harmonized with ComBat harmonization. Two-level reduction strategy with analysis of variance (ANOVA) and recursive feature elimination (RFE) was utilized sequentially to decrease high-dimensional features. Support vector machine with radial basis function kernel (RBF-SVM) was used to integrate multiscale sMRI features to construct models for early improvement prediction. Area under the curve (AUC), accuracy, sensitivity, and specificity based on the leave-one-out cross-validation (LOO-CV) and receiver operating characteristic (ROC) curve analysis were calculated to evaluate the model performance. Permutation tests were used for assessing the generalization rate.ResultsAfter 2-week ADM, 121 patients were divided into 67 ADM improvers (31 SSRIs improvers and 36 SNRIs improvers) and 54 ADM non-improvers. After two-level dimensionality reduction, 8 conventional indicators (2 VBM-based features and 6 diffusion features) and 49 radiomics features (16 VBM-based features and 33 diffusion features) were selected. The overall accuracy of RBF-SVM models based on conventional indicators and radiomics features was 74.80% and 88.19%. The radiomics model achieved the AUC, sensitivity, specificity, and accuracy of 0.889, 91.2%, 80.1% and 85.1%, 0.954, 89.2%, 87.4% and 88.5%, 0.942, 91.9%, 82.5% and 86.8% for predicting ADM improvers, SSRIs improvers and SNRIs improvers, respectively. P value of permutation tests were less than 0.001. The radiomics features predicting ADM improver were mainly located in the hippocampus, medial orbitofrontal gyrus, anterior cingulate gyrus, cerebellum (lobule vii-b), body of corpus callosum, etc. The radiomics features predicting SSRIs improver were primarily distributed in hippocampus, amygdala, inferior temporal gyrus, thalamus, cerebellum (lobule vi), fornix, cerebellar peduncle, etc. The radiomics features predicting SNRIs improver were primarily located in the medial orbitofrontal cortex, anterior cingulate gyrus, ventral striatum, corpus callosum, etc.ConclusionsThese findings suggest the radiomics analysis based on brain multiscale sMRI after ComBat harmonization could effectively predict the early improvement of ADM in adolescent MDD patients with a high accuracy, which was superior to the model based on the conventional indicators. The radiomics features with high prediction power may help for the individual selection of SSRIs and SNRIs.

Exercise repetition rate measured with simple sensors at home can be used to estimate Upper Extremity Fugl-Meyer score after stroke.

It would be valuable if home-based rehabilitation training technologies could automatically assess arm impairment after stroke. Here, we tested whether a simple measure-the repetition rate (or "rep rate") when performing specific exercises as measured with simple sensors-can be used to estimate Upper Extremity Fugl-Meyer (UEFM) score. 41 individuals with arm impairment after stroke performed 12 sensor-guided exercises under therapist supervision using a commercial sensor system comprised of two pucks that use force and motion sensing to measure the start and end of each exercise repetition. 14 of these participants then used the system at home for three weeks. Using linear regression, UEFM score was well estimated using the rep rate of one forward-reaching exercise from the set of 12 exercises (r2 = 0.75); this exercise required participants to alternately tap pucks spaced about 20 cm apart (one proximal, one distal) on a table in front of them. UEFM score was even better predicted using an exponential model and forward-reaching rep rate (Leave One Out Cross Validation (LOOCV) r2 = 0.83). We also tested the ability of a nonlinear, multivariate model (a regression tree) to predict UEFM, but such a model did not improve prediction (LOOCV r2 = 0.72). However, the optimal decision tree also used the forward-reaching task along with a pinch grip task to subdivide more and less impaired patients in a way consistent with clinical intuition. At home, rep rate for the forward-reaching exercise well predicted UEFM score using an exponential model (LOOCV r2 = 0.69), but only after we re-estimated coefficients using the home data. These results show how a simple measure-exercise rep rate measured with simple sensors-can be used to infer an arm impairment score and suggest that prediction models should be tuned separately for the clinic and home environments.

Aerobic Fitness as an Important Moderator Risk Factor for Loneliness in Physically Trained Older People: An Explanatory Case Study Using Machine Learning.

Loneliness in older people seems to have emerged as an increasingly prevalent social problem. To apply a machine learning (ML) algorithm to the task of understanding the influence of sociodemographic variables, physical fitness, physical activity levels (PAL), and sedentary behavior (SB) on the loneliness feelings of physically trained older people. The UCLA loneliness scale was used to evaluate loneliness, the Functional Fitness Test Battery was used to evaluate the correlation of sociodemographic variables, physical fitness, PAL, and SB in the loneliness feelings scores of 23 trained older people (19 women and 4 men). For this purpose, a naive Bayes ML algorithm was applied. After analysis, we inferred that aerobic fitness (AF), hand grip strength (HG), and upper limb strength (ULS) comprised the most relevant variables panel to cause high participant loneliness with 100% accuracy and F-1 score. The naive Bayes algorithm with leave-one-out cross-validation (LOOCV) predicted loneliness in trained older with a high precision. In addition, AF was the most potent variable in reducing loneliness risk.

KATZNCP: a miRNA–disease association prediction model integrating KATZ algorithm and network consistency projection

BackgroundClinical studies have shown that miRNAs are closely related to human health. The study of potential associations between miRNAs and diseases will contribute to a profound understanding of the mechanism of disease development, as well as human disease prevention and treatment. MiRNA–disease associations predicted by computational methods are the best complement to biological experiments.ResultsIn this research, a federated computational model KATZNCP was proposed on the basis of the KATZ algorithm and network consistency projection to infer the potential miRNA–disease associations. In KATZNCP, a heterogeneous network was initially constructed by integrating the known miRNA–disease association, integrated miRNA similarities, and integrated disease similarities; then, the KATZ algorithm was implemented in the heterogeneous network to obtain the estimated miRNA–disease prediction scores. Finally, the precise scores were obtained by the network consistency projection method as the final prediction results. KATZNCP achieved the reliable predictive performance in leave-one-out cross-validation (LOOCV) with an AUC value of 0.9325, which was better than the state-of-the-art comparable algorithms. Furthermore, case studies of lung neoplasms and esophageal neoplasms demonstrated the excellent predictive performance of KATZNCP.ConclusionA new computational model KATZNCP was proposed for predicting potential miRNA–drug associations based on KATZ and network consistency projections, which can effectively predict the potential miRNA–disease interactions. Therefore, KATZNCP can be used to provide guidance for future experiments.

Illicit Drug Analysis in Blood Samples with Multivariate Analysis Using Surface-Enhanced Raman Spectroscopy

This study aims to discriminate different types of illicit drugs (MDMA and THC) in blood samples using surface-enhanced Raman spectroscopy (SERS) combined with chemometric techniques including principal components analysis (PCA) and partial least squares discriminant analysis (PLS-DA). A PLS-DA classification model was built using a training data set containing Raman spectra from control and experimental groups (drug-detected blood). PLS-DA was performed for discrimination and classification among blood samples. The scores obtained in the PLS-DA model were used to evaluate the performance of the created model. The leave one out cross-validation (LOOCV) method was used for calibration and validation of the PLS-DA model. In the study, it was observed that the SERS method and chemometric techniques together could be used in drug analysis, even at low concentrations in complex body fluids such as blood. As a result, Raman spectroscopy with PCA and PLS-DA methods of data analysis could be used extensively to build similar or different classification models.

Impact of breast density on classification of infrared spectroscopy for breath-based breast cancer screening.

e13558 Background: While mammography is currently the standard of care for breast cancer screening, dense breast tissue can significantly degrade results. Alternatively, infrared spectroscopy analysis of breath offers a highly sensitive method for identifying exhaled volatile organic compounds (VOCs), which may circumvent issues with breast density. Methods: Alveolar breath samples were collected onto Tenax TA sorbent tubes using a SohnoXB™ breath sampler. Using four desorb temperatures (75, 150, 225 and 300°C), absorption spectra were measured by infrared cavity ring-down spectroscopy (IR-CRDS), a technique for measuring absorption coefficients due to VOCs in exhaled breath. Missing values in the absorption spectra were backfilled using interpolation, and the spectrum was min-max normalized and quadratic detrended. First and second derivatives of the preprocessed absorption spectra were used as features for a support vector machine machine-learning model. Features were ranked based on minimum redundancy maximum relevance (mRMR). The top 20 ranked features were selected to limit the potential for overfitting and then optimized. Model performance was validated using non-nested leave-one-out cross-validation (LOOCV) and nested LOOCV to provide optimistic and pessimistic results, respectively. Results: Absorption spectra from 111 participants (71 positive, 40 control) were used. Of the positive subjects, 30 had low- and 31 had high-density breast tissue (measures missing for 10). Model performance is outlined. A subgroup analysis compared model performance for subjects with low- and high-density breast tissue for both non-nested and nested LOOCV models. Breast density data was not captured for control subjects thus they were not included in the subgroup analysis. Fisher’s exact test was performed to assess for significant difference between model performance for those with low- vs high-density breast tissue, resulting in a p-value of 1.00. Conclusions: Our results suggest that the classification of alveolar breath using IR-CRDS is a promising technique for the detection of breast cancer that is independent of breast density. Breath analysis may therefore become a new alternative or corroborative to mammography to support clinical decision-making. [Table: see text]

A motion model-guided 4D dose reconstruction for pencil beam scanned proton therapy

Objective. 4D dose reconstruction in proton therapy with pencil beam scanning (PBS) typically relies on a single pre-treatment 4DCT (p4DCT). However, breathing motion during the fractionated treatment can vary considerably in both amplitude and frequency. We present a novel 4D dose reconstruction method combining delivery log files with patient-specific motion models, to account for the dosimetric effect of intra- and inter-fractional breathing variability. Approach. Correlation between an external breathing surrogate and anatomical deformations of the p4DCT is established using principal component analysis. Using motion trajectories of a surface marker acquired during the dose delivery by an optical tracking system, deformable motion fields are retrospectively reconstructed and used to generate time-resolved synthetic 4DCTs (‘5DCTs’) by warping a reference CT. For three abdominal/thoracic patients, treated with respiratory gating and rescanning, example fraction doses were reconstructed using the resulting 5DCTs and delivery log files. The motion model was validated beforehand using leave-one-out cross-validation (LOOCV) with subsequent 4D dose evaluations. Moreover, besides fractional motion, fractional anatomical changes were incorporated as proof of concept. Main results. For motion model validation, the comparison of 4D dose distributions for the original 4DCT and predicted LOOCV resulted in 3%/3 mm gamma pass rates above 96.2%. Prospective gating simulations on the p4DCT can overestimate the target dose coverage V95% by up to 2.1% compared to 4D dose reconstruction based on observed surrogate trajectories. Nevertheless, for the studied clinical cases treated with respiratory-gating and rescanning, an acceptable target coverage was maintained with V95% remaining above 98.8% for all studied fractions. For these gated treatments, larger dosimetric differences occurred due to CT changes than due to breathing variations. Significance. To gain a better estimate of the delivered dose, a retrospective 4D dose reconstruction workflow based on motion data acquired during PBS proton treatments was implemented and validated, thus considering both intra- and inter-fractional motion and anatomy changes.

Recognition of Children’s Facial Expressions Using Deep Learned Features

The emotional well-being of a child is crucial for their successful integration into society as a productive individual. While technology has made significant strides in enabling machines to decipher human emotional signals, current research in emotion recognition primarily prioritizes adults, disregarding the fact that children develop emotional awareness at an early stage. This highlights the need to explore how machines can recognize facial expressions in children, although the absence of a standardized database poses a challenge. In this study, we propose a system that employs Convolutional-Neural-Network (CNN)-based models, such as VGG19, VGG16, and Resnet50, as feature extractors, and Support Vector Machine (SVM) and Decision Tree (DT) for classification, to automatically recognize children’s expressions using a video dataset, namely Children’s Spontaneous Facial Expressions (LIRIS-CSE). Our system is evaluated through various experimental setups, including 80–20% split, K-Fold Cross-Validation (K-Fold CV), and leave one out cross-validation (LOOCV), for both image-based and video-based classification. Remarkably, our research achieves a promising classification accuracy of 99% for image-based classification, utilizing features from all three networks with SVM using 80–20% split and K-Fold CV. For video-based classification, we achieve 94% accuracy using features from VGG19 with SVM using LOOCV. These results surpass the performance of the original work, which reported an average image-based classification accuracy of 75% on their LIRIS-CSE dataset. The favorable outcomes obtained from our research can pave the way for the practical application of our proposed emotion recognition methodology in real-world scenarios.

Nanomotion technology in combination with machine learning: a new approach for a rapid antibiotic susceptibility test for Mycobacterium tuberculosis

Nanomotion technology is a growth-independent approach that can be used to detect and record the vibrations of bacteria attached to cantilevers. We have developed a nanomotion-based antibiotic susceptibility test (AST) protocol for Mycobacterium tuberculosis (MTB). The protocol was used to predict strain phenotype towards isoniazid (INH) and rifampicin (RIF) using a leave-one-out cross-validation (LOOCV) and machine learning techniques. This MTB-nanomotion protocol takes 21 h, including cell suspension preparation, optimized bacterial attachment to functionalized cantilever, and nanomotion recording before and after antibiotic exposure. We applied this protocol to MTB isolates (n = 40) and were able to discriminate between susceptible and resistant strains for INH and RIF with a maximum sensitivity of 97.4% and 100%, respectively, and a maximum specificity of 100% for both antibiotics when considering each nanomotion recording to be a distinct experiment. Grouping recordings as triplicates based on source isolate improved sensitivity and specificity to 100% for both antibiotics. Nanomotion technology can potentially reduce time-to-result significantly compared to the days and weeks currently needed for current phenotypic ASTs for MTB. It can further be extended to other anti-TB drugs to help guide more effective TB treatment.

Distinguishing Patients with MRI-Negative Temporal Lobe Epilepsy from Normal Controls Based on Individual Morphological Brain Network.

Temporal Lobe Epilepsy (TLE) is the most common subtype of focal epilepsy and the most refractory to drug treatment. Roughly 30% of patients do not have easily identifiable structural abnormalities. In other words, MRI-negative TLE has normal MRI scans on visual inspection. Thus, MRI-negative TLE is a diagnostic and therapeutic challenge. In this study, we investigate the cortical morphological brain network to identify MRI-negative TLE. The 210 cortical ROIs based on the Brainnetome atlas were used to define the network nodes. The least absolute shrinkage and selection operator (LASSO) algorithm and Pearson correlation methods were used to calculate the inter-regional morphometric features vector correlation respectively. As a result, two types of networks were constructed. The topological characteristics of networks were calculated by graph theory. Then after, a two-stage feature selection strategy, including a two-sample t-test and support vector machine-based recursive feature elimination (SVM-RFE), was performed in feature selection. Finally, classification with support vector machine (SVM) and leave-one-out cross-validation (LOOCV) was employed for the training and evaluation of the classifiers. The performance of two constructed brain networks was compared in MRI-negative TLE classification. The results indicated that the LASSO algorithm achieved better performance than the Pearson pairwise correlation method. The LASSO algorithm provides a robust method of individual morphological network construction for distinguishing patients with MRI-negative TLE from normal controls.

A simple procedure for a national wheat yield forecast

National crop yields are difficult to estimate during a crop season and are usually only known after crop harvest. The goal of this study was to develop a simple methodology to estimate national wheat yields that could be easily applied to any country and crop. Twenty years of readily available global gridded monthly climate data (0.5°) across wheat cultivated areas of a country were correlated with national wheat production-weighted mean climate indices to determine the single most representative climate grid cell for the entire wheat region. The same 20 years of monthly climate data from this most representative grid cell were then used to build statistical models to estimate trend-corrected national wheat yields, including a Stepwise Regression function (Stepwise) set with the Bayesian information criterion (BIC), a least absolute shrinkage and selection operator algorithm (Lasso), and a Random Forest machine-learning algorithm (Random Forest). The best of the three models estimated trend-corrected national yield variability from 1998 to 2021 for Brazil with an rRMSE of 9.1%. In an additional validation, the same approach was then applied to national wheat yields in France and Russia resulting in an rRMSE from a Leave One Out Cross Validation (LOOCV) of 6.7% and 6.4%, respectively. As the statistical models employed monthly climate data from within a season, national yield predictions are possible during a cropping season before crop harvest by using the best performing model with the predictability of a national yield further improving towards harvest. This approach should be applicable to any crop and nation.

Distinguishing Parkinson's Disease with GLCM Features from the Hankelization of EEG Signals.

This study proposes a novel method that uses electroencephalography (EEG) signals to classify Parkinson's Disease (PD) and demographically matched healthy control groups. The method utilizes the reduced beta activity and amplitude decrease in EEG signals that are associated with PD. The study involved 61 PD patients and 61 demographically matched controls groups, and EEG signals were recorded in various conditions (eyes closed, eyes open, eyes both open and closed, on-drug, off-drug) from three publicly available EEG data sources (New Mexico, Iowa, and Turku). The preprocessed EEG signals were classified using features obtained from gray-level co-occurrence matrix (GLCM) features through the Hankelization of EEG signals. The performance of classifiers with these novel features was evaluated using extensive cross-validations (CV) and leave-one-out cross-validation (LOOCV) schemes. This method under 10 × 10 fold CV, the method was able to differentiate PD groups from healthy control groups using a support vector machine (SVM) with an accuracy of 92.4 ± 0.01, 85.7 ± 0.02, and 77.1 ± 0.06 for New Mexico, Iowa, and Turku datasets, respectively. After a head-to-head comparison with state-of-the-art methods, this study showed an increase in the classification of PD and controls.

A New Method for Spatial Estimation of Water Quality Using an Optimal Virtual Sensor Network and In Situ Observations: A Case Study of Chemical Oxygen Demand.

Accurate water quality estimation is important for water environment monitoring and water resource management and has emerged as a pivotal aspect of ecological rehabilitation and sustainable development. However, due to the strong spatial heterogeneity of water quality parameters, it is still challenging to obtain highly accurate spatial patterns of them. Taking chemical oxygen demand as an example, this study proposes a novel estimation method for generating highly accurate chemical oxygen demand fields in Poyang Lake. Specifically, based on the different water levels and monitoring sites in Poyang Lake, an optimal virtual sensor network was first established. A Taylor expansion-based method with integration of spatial correlation and spatial heterogeneity was developed by considering environmental factors, the optimal virtual sensor network, and existing monitoring stations. The proposed approach was evaluated and compared with other approaches using a leave-one cross-validation process. Results show that the proposed method exhibits good performance in estimating chemical oxygen demand fields in Poyang Lake, with mean absolute error improved by 8% and 33%, respectively, on average, when compared with classical interpolators and remote sensing methods. In addition, the applications of virtual sensors improve the performance of the proposed method, with mean absolute error and root mean squared error values reduced by 20% to 60% over 12 months. The proposed method provides an effective tool for estimating highly accurate spatial fields of chemical oxygen demand concentrations and could be applied to other water quality parameters.

A gravity inspired clustering algorithm for gene selection from high-dimensional microarray data

ABSTRACT The problem of selecting a subset of genes relevant to a particular disease classification task from high-dimensional gene microarray data has been an area of active research in the scientific community. A wide variety of methods ranging from assigning scores to genes based on their interactions with the output to the Gravity inspired clustering algorithm with the hybrid approach have been proposed. The feature sets selected by this approach have been evaluated by using the Support Vector Machine (SVM) with Leave One Out Cross Validation (LOOCV). The method has been evaluated on twelve different publicly available microarray cancer datasets and the proposed method provided good accuracy of 90.16% to the colon dataset, 98.59% to the ALL-AML dataset, 97.03% to the lung cancer dataset, 93.07% to the prostate cancer dataset, 98.41% to the meningioma dataset, 83.72% to the brain tumour dataset and has been shown to compete with existing methods.