Leave-one-out Cross-validation Research Articles

Mapping of Chl-a in Indonesia’s Five Priority Lakes uSing Linear Regression (Case Studies: Lake Singkarak, Toba, Maninjau, Matano, and Towuti)

In Indonesia, 15 priority lakes need to be monitored regularly. Model algorithm development is the answer to accelerating lake water quality monitoring through chlorophyll-a concentration monitoring. This research aims to evaluate, develop, and find the most accurate global model algorithm for chlorophyll-a concentration mapping on Lake Singkarak, Toba, Maninjau, Matano, and Towuti. Algorithm development was made based on the existing 22-band ratio algorithm. Leave One Out Cross Validation (LOOCV) was used to evaluate the performance of all tested algorithms through the R2, NMAE, and RMSE accuracy results using 42 in-situ sample points of chlorophyll-a collected on five lakes. Lake Singkarak has an RMSE of 0.1 mg/m3, NMAE of 29.9%, R of 0.7, and R2 of 1.0. Lake Toba in the dry season has RMSE of 0.7 mg/m3, NMAE of 22.5%, R of 0.7 and R2 of 0.3. In the rainy season, Lake Toba has an RMSE of 0.1 mg/m3, NMAE of 7.1%, R of 0.5, and R2 of 0.3. Lake Maninjau has an RMSE of 0.6 mg/m3, NMAE of 32.5%, R of 0.5, and R2 of 0.3. Lake Matano has RMSE of 0.0 mg/m3, NMAE of 4.2%, R of 1.0 and R2 of 1.0. Lake Towuti has an RMSE of 0.0 mg/m3, NMAE of 3.4%, R of 1.0, and R2 of 1.0. The most accurate algorithm for Lake Singkarak, Toba (dry season), Toba (rainy season), Maninjau, Matano, and Towuti are C6, C6, C2, A2, C10, and B3, respectively

Supervised and unsupervised machine learning approaches for prediction and geographical discrimination of Iranian saffron ecotypes based on flower-related and phytochemical attributes

A two-year field experiment (2014–2016; Zanjan, Iran) was conducted to monitor potential diversity pattern and adaptability power among 18 Iranian saffron ecotypes under Zanjan climatological conditions using seven flower-related and three qualitative traits (crocin, picrocrocin, and safranal, determined by UV–visible spectra), and analyzed by supervised and unsupervised approaches. A range of variability was recorded among the ecotypes, and despite some exceptions, overall, saffron corms produced higher amounts of studied features across the second year. The Feizabad ecotype was recommended to acquire maximum qualitative criteria (category I; based on ISO Normative 3632 grading system), while for flower-related parameters several ecotypes (e.g., Ghaien, Bardeskan, Torbat-Jam, and Gonabad) could be applied for Zanjan climatological conditions. Based on the results of Leave-One-Out Cross-Validation (LOOCV), various prediction values were computed for all 10 classifiers of LDA, QDA, FDA, MDA, RDA, Naive Bayes, Decision Tree, Linear SVM, Radial SVM, and Random Forest in terms of accuracy, sensitivity and specificity parameters. Among which, Random Forest and LDA with the values of 0.91 and 0.78 possessed the highest and the lowest amounts of accuracy, respectively. Finally, considering the highest accuracy value of the superior classification model of Random forest, both feature subsets of “FFW, FDW, Picrocrocin, Safranal, and Crocin” and “SFW, FDW, Picrocrocin, Safranal, and Crocin” were nominated as the most powerful elements (comparing to the remaining 1021 feature subsets) to make accurate discrimination between Khorasan and non-Khorasan saffron ecotypes. The results, overall, revealed that saffron ecotypes followed different responses under Zanjan climatological circumstances, and Random Forest is more suitable for accurately predicting saffron corms from different provenances.

Comparative QSAR Modeling for Predicting Anticancer Potency of Imidazo[4,5-b]Pyridine Derivatives Using GA-MLR and BP-ANN Techniques

Background: Prediction of toxicity of imidazo[4,5-b]pyridine derivatives is carried out using GA-MLR and BPANN methods. Objective: A quantitative structure-property relationship (QSPR) was determined based on methods, including genetic algorithm-multiple linear regression (GA-MLR) and backpropagation artificial neural network (BP-ANN). These methods were employed for modeling and predicting the anticancer potency of imidazo[4,5-b]pyridine derivatives. Materials and Methods: A dataset of imidazo[4,5-b]pyridine derivatives was randomly divided into two groups, training and test sets consisting of 75% and 25% of data points, respectively. The optimized conformation of compounds was obtained using the DFT-B3LYP method and 6-31G* basis sets level with Gaussian 09 software. A large number of molecular descriptors were calculated using Dragon software. The QSAR models were optimized using multiple linear regressions (MLR). Results: The most relevant molecular descriptors were obtained using the genetic algorithm (GA) and backward stepwise regression. The predictive powers of the GA-MLR models were studied using leaveone- out (LOO) cross-validation and an external test set. Conclusion: The obtained results of statistical parameters showed the BP-ANN model to have better performance compared to the GA-MLR model. To assess the predictive ability of QSAR models, many statistical terms, such as correlation coefficient (R2), leave-one-out cross-validation (LOOCV), root mean squared error (RMSE), and external and internal validation were used. The results of validation methods demonstrate the QSAR model to be robust and with high predictivity.

Structural sensitivity to reliability of flexible AMOLED modules using mechanical simulation and machine learning

The flexibility and durability of flexible AMOLEDs are considered mutually exclusive, and their structural integrity under severe load conditions can be attained by minimizing the trade-off between these two properties. In this regard, the thickness and elastic modulus of the plastic films comprising flexible AMOLEDs are crucial design variables that determine their reliability. This study proposes a method for predicting the performance sensitivity of an AMOLED to its flexibility and durability using mechanical simulation and machine learning. A combination of 1000 thicknesses and elastic moduli, generated by Latin hypercube sampling, was used for the mechanical simulation. The results of the mechanical simulation were used to train various machine-learning algorithms, and the performance was evaluated using leave-one-out cross-validation (LOOCV). The CatBoost algorithm, which produced the best accuracy, and Kernel Shapley Additive Explanations (SHAP), were utilized to represent the sensitivity of the thickness and elastic modulus and their mutual exclusiveness is experimentally verified. These results will provide crucial information for the optimal design of flexible AMOLED modules.

Comparison of Error Prediction Methods in Claassification Modeling with CHAID Methods for Balanced Data

Chi-Squared Automatic Interaction Detection (CHAID) is an exploratory method for classifying data by building classification trees. The classification result are displayed in the form of a tree diagram model. After the model is formed, it is necessary to calculate the accuracy of the model. The goal is to see the performance of the model. The accuracy of this model can be determined by calculating the level of prediction error in the model. The error rate prediction method works by dividing data into training data and testing data. There are three methods in the error rate prediction method, such as Leave one out cross validation (LOOCV), Hold out, and k-fold cross validation. These methods have different performance in dividing data into training data and test data, so that each method has advantages and disadvantages. Therefore, a comparison of the three error rate prediction methods was carried out with the aim of determining the appropriate method for the CHAID. This research is included in experimental research and uses simulation data from data generation results in RStudio. This comparison is carried out by considering several factors, namely the marginal probability matrix and different correlations. The comparison results will be observed using a boxplot by looking at the median error rate and lowest variance. This research found that k-fold cross validation is the most suitable error rate prediction method applied to the CHAID method for balanced data.

Sequential Monte Carlo with cross-validated neural networks for complexity of hyperbolic black hole solutions in 4D

This paper investigates the self-similar solutions of the Einstein-axion-dilaton configuration from type IIB string theory and the global SL(2,R) symmetry. We consider the Continuous Self Similarity (CSS), where the scale transformation is controlled by an SL(2, R) boost or hyperbolic translation. The solutions stay invariant under the combination of space-time dilation with internal SL(2,R) transformations. We develop a new formalism based on Sequential Monte Carlo (SMC) and artificial neural networks (NNs) to estimate the self-similar solutions to the equations of motion in the hyperbolic class in four dimensions. Due to the complex and highly nonlinear patterns, researchers typically have to use various constraints and numerical approximation methods to estimate the equations of motion; thus, they have to overlook the measurement errors in parameter estimation. Through a Bayesian framework, we incorporate measurement errors into our models to find the solutions to the hyperbolic equations of motion. It is well known that the hyperbolic class suffers from multiple solutions where the critical collapse functions have overlap domains for these solutions. To deal with this complexity, for the first time in literature on the axion-dilaton system, we propose the SMC approach to obtain the multi-modal posterior distributions. Through a probabilistic perspective, we confirm the deterministic α\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\alpha $$\\end{document} and β\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\beta $$\\end{document} solutions available in the literature and determine all possible solutions that may occur due to measurement errors. We finally proposed the penalized Leave-One-Out Cross-validation (LOOCV) to combine the Bayesian NN-based estimates optimally. The approach enables us to determine the optimum weights while dealing with the co-linearity issue in the NN-based estimates and better predict the critical functions corresponding to multiple solutions of the equations of motion.

Classifying dynamic motor imagery with the locals-balanced extreme learning machine

The dynamic motor imagery (dMI) provides additional benefits as compared to the traditional motor imagery (MI) in training and neurorehabilitation field. The objective of this work is to develop a brain-computer interface (BCI) for dMI electroencephalograph (EEG). We propose a novel method by combining synchronization likelihood (SL) based functional brain network (FBN) and modified extreme learning machine (ELM) to interpret EEG. The proposed method 1) uses L1-norm and adaptive threshold instead of L2-norm and empirical threshold in SL method; 2) improves the procedure of FBN calculations (namely, SL matrix binarization) by combining threshold and MST methods; 3) detects two defects in standard ELM in fusion, and proposes the locals-balanced ELM (LBELM); 4) employs the special version of leave-one-out cross validation (LOO) approach improved the computational complexity of optimal threshold for binarization, the most effective fusion of two complementary features, and the optimal regularization parameters for regularized LBELM. As compared with the empirical threshold, the recognition rate of the optimal threshold increases by 1.39–11.11 %. Similarly, as compared with the regularized ELM, the recognition rate of LBELM increases by 2.78–9.73 %.

Analysis of treatment response based on 1.5T magnetic resonance imaging texture analysis in stereotactic body radiotherapy of hepatocellular carcinoma

PurposeMagnetic Resonance Image (MRI)-guided stereotactic body radiotherapy (MRgSBRT) is a promising technique in the treatment of hepatocellular carcinoma (HCC). However, treatment response varies among patients. The aim of this study was to evaluate the treatment response using radiomics features extracted from 1.5 T MRI in HCC patients treated with SBRT. Materials and methods19 patients with biopsy-proven HCC who were treated with SBRT by 1.5 T MRI-guided were enrolled, all of whom were treated with 8–10 fractions with biological effective dose (BED) range of 95.2–107.1 Gy. We acquired images of pretreatment, delivered BEDs of 35–40 Gy, and delivered BEDs of 55–60 Gy. We combined three classic feature selection methods: least absolute shrinkage and selection operator (LASSO), extreme gradient boosting (XGBoost), and random forest (RF) to extract the intersection of radiomics features from the gross tumor volume (GTV) and features were averaged over the three fractions. We used the receiver operating characteristic (ROC) area under curve (AUC) obtained using leave-one-out cross-validation (LOOCV) to assess the predictive capacity. ResultsSeven patients showed response to treatment based on post-treatment imaging studies. The optimum intersection of radiomics features selected by three methods was Neighborhood Grey Tone Difference Matrix (NGTDM) contrast and NGTDM strength. The logistic regression-based model achieved an AUC of 0.821 (95% confidence interval, 0.618–1). ConclusionsRadiomics features containing biological prognostic information extracted during MRgSBRT could predict tumor treatment response, facilitating stratification of high-risk patients and providing clinical application value for individualized care of patients with different response to treatment.

Bringing 2D Eclipse Mapping out of the Shadows with Leave-one-out Cross Validation

Eclipse mapping is a technique for inferring 2D brightness maps of transiting exoplanets from the shape of an eclipse light curve. With JWST’s unmatched precision, eclipse mapping is now possible for a large number of exoplanets. However, eclipse mapping has only been applied to two planets, and the nuances of fitting eclipse maps are not yet fully understood. Here, we use Leave-one-out Cross Validation (LOO-CV) to investigate eclipse mapping, with application to a JWST NIRISS/SOSS observation of the ultrahot Jupiter WASP-18b. LOO-CV is a technique that provides insight into the out-of-sample predictive power of models on a data-point-by-data-point basis. We show that constraints on planetary brightness patterns behave as expected, with large-scale variations driven by the phase-curve variation in the light curve and smaller-scale structures constrained by the eclipse ingress and egress. For WASP-18b we show that the need for higher model complexity (smaller-scale features) is driven exclusively by the shape of the eclipse ingress and egress. We use LOO-CV to investigate the relationship between planetary brightness map components when mapping under a positive-flux constraint to better understand the need for complex models. Finally, we use LOO-CV to understand the degeneracy between the competing “hot spot” and “plateau” brightness map models of WASP-18b, showing that the plateau model is driven by the ingress shape and the hot spot model is driven by the egress shape, but preference for neither model is due to outliers or unmodeled signals. Based on this analysis, we make recommendations for the use of LOO-CV in future eclipse-mapping studies.

Hip and lumbosacral joint centre locations in asian population: Biases produced by existing regression equations and development of new equations

The hip and lumbosacral joint centre (HJC and LSJC) predictions are required to analyse the lumbo-pelvic-hip dynamics during various human motions. Some HJC and LSJC regression equations based on pelvic dimension have been developed; however, the pre-existing methods need to be re-evaluated, and methodological reconsideration may improve the regression methods. Here we show that pre-existing methods produce biased predictions of the LSJC and HJC in 23 male and 24 female Japanese adults, and that the biases in the LSJC differ between sexes, using magnetic resonance imaging (MRI) around the pelvis. Compared with directly measured locations on MRI, the pre-existing regression equations predict LSJC to be more posterior in males and more inferior and posterior in females, and HJC to be more medial in both sexes. The better pre-existing regression equation for LSJC height differs between sexes, with pelvic-width-base better in males and pelvic-depth-base better in females, respectively. We suggest the unsuitability of pre-existing methods to our dataset consisting of Japanese adults and the importance of considering sex differences in regression methods. We propose regression equations to predict HJC and LSJC, considering soft-tissue thickness, sex differences, and a height-directional measure, using least absolute shrinkage and selection operator regression. We validate them using leave-one-out cross-validation (LOOCV). LOOCV shows that our model produces negligible biases and smaller absolute errors than the pre-existing regressions; in particular, the anteroposterior absolute error for LSJC is less than half that of the pre-existing regression. Our regression equation can be a powerful solution for accurate motion analysis.

Evaluation of relative pollen productivities in temperate China for reliable pollen-based quantitative reconstructions of Holocene plant cover.

The Landscape Reconstruction Algorithm (LRA) is regarded as the soundest approach for quantifying taxon-specific plant cover from pollen data. The reliability of relative pollen productivity (RPP) estimates is fundamental in the accuracy of quantitative vegetation reconstruction using the LRA approach. Inconsistent RPP estimates produced by different studies can cast doubt on the reliability and applicability of quantitative vegetation reconstruction. Therefore, it is crucial that the RPP estimates are evaluated before being applied for quantitative vegetation reconstruction. We have tested two alternative approaches, namely, a leave-one-out cross-validation (LOO) method and a splitting-by-subregion strategy, using surface pollen assemblages and the REVEALS model-the first step in the LRA-to evaluate the reliability of RPPs estimates of 10 target taxa obtained in the cultural landscape of Shandong. We compared the REVEALS estimates (RVs) with observations of regional vegetation abundance (OBVs) and pollen proportions (PPs). The RVs of all taxa are generally closer to OBVs than PPs, and the degree of similarity depends strongly on the abundance of individual taxa in plant and pollen; taxa dominant in the region show the highest similarity between RVs and OBVs, such as Artemisia, Poaceae, and Humulus. The RVs of all herb taxa except Humulus and Asteraceae SF Cichorioideae are slightly overrepresented, and the RVs of all tree taxa are underrepresented except for Castanea. The comparison of RVs with OBVs collected from different spatial extents shows that the RVs of all herb taxa are more similar to OBVs collected from shorter distances (100km and 75km for the entire region and the subregion, respectively), whereas the RVs of all tree taxa are more similar to OBVs collected from longer distances (150km and 100km for the entire region and the subregion, respectively). Furthermore, our findings highlight the importance to test different sizes of area for vegetation surveys for evaluation of the RVs given that the appropriate size of vegetation survey may vary between low pollen producers (mainly herbs) and high pollen producers (mainly trees). We consider that the LOO strategy is the best approach in this case study for evaluating the RPP estimates from surface moss polsters. The evaluation confirms the reliability of the obtained RPP estimates for their potential application in quantitative reconstruction of vegetation abundance in temperate China.

Total and component forest aboveground biomass inversion via LiDAR-derived features and machine learning algorithms

Forest aboveground biomass (AGB) and its biomass components are key indicators for assessing forest ecosystem health, productivity, and carbon stocks. Light Detection and Ranging (LiDAR) technology has great advantages in acquiring the vertical structure of forests and the spatial distribution characteristics of vegetation. In this study, the 56 features extracted from airborne LiDAR point cloud data were used to estimate forest total and component AGB. Variable importance–in–projection values calculated through a partial least squares regression algorithm were utilized for LiDAR-derived feature ranking and optimization. Both leave-one-out cross-validation (LOOCV) and cross-validation methods were applied for validation of the estimated results. The results showed that four cumulative height percentiles (AIH30,AIH40, AIH20, and AIH25), two height percentiles (H8 and H6), and four height-related variables (Hmean, Hsqrt, Hmad, and Hcurt) are ranked more frequently in the top 10 sensitive features for total and component forest AGB retrievals. Best performance was acquired by random forest (RF) algorithm, with R2 = 0.75, root mean square error (RMSE) = 22.93 Mg/ha, relative RMSE (rRMSE) = 25.30%, and mean absolute error (MAE) = 19.26 Mg/ha validated by the LOOCV method. For cross-validation results, R2 is 0.67, RMSE is 24.56 Mg/ha, and rRMSE is 25.67%. The performance of support vector regression (SVR) for total AGB estimation is R2 = 0.66, RMSE = 26.75 Mg/ha, rRMSE = 28.62%, and MAE = 22.00 Mg/ha using LOOCV validation and R2 = 0.56, RMSE = 30.88 Mg/ha, and rRMSE = 31.41% by cross-validation. For the component AGB estimation, the accuracy from both RF and SVR algorithms was arranged as stem > bark > branch > leaf. The results confirmed the sensitivity of LiDAR-derived features to forest total and component AGBs. They also demonstrated the worse performance of these features for retrieval of leaf component AGB. RF outperformed SVR for both total and component AGB estimation, the validation difference from LOOCV and cross-validation is less than 5% for both total and component AGB estimated results.

Prediction of Radiation Treatment Response for Locally Advanced Rectal Cancer via a Longitudinal Trend Analysis Framework on Cone-Beam CT.

Locally advanced rectal cancer (LARC) presents a significant challenge in terms of treatment management, particularly with regards to identifying patients who are likely to respond to radiation therapy (RT) at an individualized level. Patients respond to the same radiation treatment course differently due to inter- and intra-patient variability in radiosensitivity. In-room volumetric cone-beam computed tomography (CBCT) is widely used to ensure proper alignment, but also allows us to assess tumor response during the treatment course. In this work, we proposed a longitudinal radiomic trend (LRT) framework for accurate and robust treatment response assessment using daily CBCT scans for early detection of patient response. The LRT framework consists of four modules: (1) Automated registration and evaluation of CBCT scans to planning CT; (2) Feature extraction and normalization; (3) Longitudinal trending analyses; and (4) Feature reduction and model creation. The effectiveness of the framework was validated via leave-one-out cross-validation (LOOCV), using a total of 840 CBCT scans for a retrospective cohort of LARC patients. The trending model demonstrates significant differences between the responder vs. non-responder groups with an Area Under the Curve (AUC) of 0.98, which allows for systematic monitoring and early prediction of patient response during the RT treatment course for potential adaptive management.

Bayesian workflow for the investigation of hierarchical classification models from tau-PET and structural MRI data across the Alzheimer's disease spectrum.

Alzheimer's disease (AD) diagnosis in its early stages remains difficult with current diagnostic approaches. Though tau neurofibrillary tangles (NFTs) generally follow the stereotypical pattern described by the Braak staging scheme, the network degeneration hypothesis (NDH) has suggested that NFTs spread selectively along functional networks of the brain. To evaluate this, we implemented a Bayesian workflow to develop hierarchical multinomial logistic regression models with increasing levels of complexity of the brain from tau-PET and structural MRI data to investigate whether it is beneficial to incorporate network-level information into an ROI-based predictive model for the presence/absence of AD. This study included data from the Translational Biomarkers in Aging and Dementia (TRIAD) longitudinal cohort from McGill University's Research Centre for Studies in Aging (MCSA). Baseline and 1 year follow-up structural MRI and [18F]MK-6240 tau-PET scans were acquired for 72 cognitive normal (CN), 23 mild cognitive impairment (MCI), and 18 Alzheimer's disease dementia subjects. We constructed the four following hierarchical Bayesian models in order of increasing complexity: (Model 1) a complete-pooling model with observations, (Model 2) a partial-pooling model with observations clustered within ROIs, (Model 3) a partial-pooling model with observations clustered within functional networks, and (Model 4) a partial-pooling model with observations clustered within ROIs that are also clustered within functional brain networks. We then investigated which of the models had better predictive performance given tau-PET or structural MRI data as an input, in the form of a relative annualized rate of change. The Bayesian leave-one-out cross-validation (LOO-CV) estimate of the expected log pointwise predictive density (ELPD) results indicated that models 3 and 4 were substantially better than other models for both tau-PET and structural MRI inputs. For tau-PET data, model 3 was slightly better than 4 with an absolute difference in ELPD of 3.10 ± 1.30. For structural MRI data, model 4 was considerably better than other models with an absolute difference in ELPD of 29.83 ± 7.55 relative to model 3, the second-best model. Our results suggest that representing the data generating process in terms of a hierarchical model that encompasses both ROI-level and network-level heterogeneity leads to better predictive ability for both tau-PET and structural MRI inputs over all other model iterations.

Estimating dim light melatonin onset time in children using delta changes in melatonin.

We aimed to establish a method for estimating dim light melatonin onset (DLMO) using mathematical slopes calculated from melatonin concentrations at three sampling points before and after sleep in children. The saliva of 30 children (mean age ± SD: 10.2 ± 1.3years old) was collected under dim-light conditions up to six times every hour starting at 17:30 (t17), namely, 18:30 (t18), 19:30 (t19), 20:30 (t20), 21:30 (t21), 22:30 (t22), and 23:30 (t23), in the evening, and at 6:00 (t30) the following morning. We calculated (mathematical slope between melatonin concentrations at t18 and t20, t21 or t22), (the slope between t20, t21 or t22, and t30), and , which is generated by subtracting from . DLMO was estimated by multiple regression analysis with the leave-one-out cross-validation (LOOCV) method using and , and . The intraclass correlation coefficient (ICC) between the estimated and measured DLMOs was used as the index of estimation accuracy. DLMOs estimated using multiple regression equations with and yielded significant ICCs for the measured DLMOs, with the largest ICC at t20 (ICC = 0.634). Additionally, the ICC between the estimated and measured DLMOs using the equation with was significant, with a larger ICC at t20 (ICC = 0.726) than that of the equation with and . Our results showed that DLMO could be estimated with a certain level of accuracy from salivary melatonin levels at three time points before and after sleep in children. The online version contains supplementary material available at 10.1007/s41105-023-00493-x.

A robust seizure detection and prediction method with feature selection and spatio-temporal casual neural network model.

Epilepsy is a fairly common condition that affects the brain and causes frequent seizures. The sudden and recurring epilepsy brings a series of safety hazards to patients, which seriously affects the quality of their life. Therefore, real-time diagnosis of Electroencephalogram (EEG) in epilepsy patients is of great significance. However, the conventional methods take in a tremendous amount of features to train the models, resulting in high computation cost and low portability. Our objective is to propose an efficient, light and robust seizure detecting and predicting algorithm. The algorithm is based on an interpretative feature selection method and Spatial-Temporal Causal Neural Network (STCNN). The feature selection method eliminates the interference factors between different features and reduces the model size and training difficulties. The STCNN model takes both temporal and spatial information to accurately and dynamically track and diagnose the changing of the features. Considering the differences between medical application scenarios and patients, Leave-One-Out Cross Validation (LOOCV) and Cross-Patient Validation (CPV) methods are used to conduct experiments on the CHB-MIT, Siena and Kaggle competition datasets. In LOOCV-based method, the detection accuracy and prediction sensitivity have been improved. A significant improvement is also achieved in the CPV-based method. The experimental results show that our proposed algorithm exhibits superior performance and robustness in seizure detection and prediction, which indicates it has higher capability to deal with different and complicated clinical situations.

The influence of agricultural land on the level of airborne Alternaria spores

Aim of the research was to investigate the impact of agricultural activity on the concentration of Alternaria spores.
 Materials and methods. The study was carried out at the Department of Medical Biology, Parasitology and Genetics of the ZSMPhU. Samples were collected using a 7-day volumetric sampler of the Hirst type, using the volumetric method. Samples were identified under a light microscope, and spore identification and counting were limited to genus levels. The relationship between seasonal Alternaria spore levels and harvest rates was analysed using Pearson's correlation method. The effect of meteorological conditions and agricultural activity on the daily concentration of Alternaria was analysed using stepwise correlation based on logarithmically transformed daily average spore counts. Classical leave-one-out cross-validation (LOOCV) was used to estimate the mean square error (MSE), associated with this model and Bayesian information criterion (BIC) was used to assess its accuracy.
 Results. Seasonal characteristics of Alternaria spores and agricultural activity in Zaporizhzhia and Dnipro regions were analysed. The connection of some seasonal and daily indicators with harvesting rates and meteorological conditions was determined. Two models with 5 and 9 parameters were found that best explain the dynamics of Alternaria spores.
 Conclusions. The most significant parameters positively correlated with Alternaria spore levels were temperature, pressure, westerly wind and wheat yield; relative humidity was negatively correlated

Simultaneous diagnosis of tuberculous pleurisy and malignant pleural effusion using metagenomic next-generation sequencing (mNGS)

BackgroundMetagenomic next-generation sequencing (mNGS) has become a powerful tool for pathogen detection, but the value of human sequencing reads generated from it is underestimated.MethodsA total of 138 patients with pleural effusion (PE) were diagnosed with tuberculous pleurisy (TBP, N = 82), malignant pleural effusion (MPE, N = 35), or non-TB infection (N = 21), whose PE samples all underwent mNGS analysis. Clinical TB tests including culture, Acid-Fast Bacillus (AFB) test, Xpert, and T-SPOT, were performed. To utilize mNGS for MPE identification, 25 non-MPE samples (20 TBP and 5 non-TB infection) were randomly selected to set human chromosome copy number baseline and generalized linear modeling was performed using copy number variant (CNV) features of the rest 113 samples (35 MPE and 78 non-MPE).ResultsThe performance of TB detection was compared among five methods. T-SPOT demonstrated the highest sensitivity (61% vs. culture 32%, AFB 12%, Xpert 35%, and mNGS 49%) but with the highest false-positive rate (10%) as well. In contrast, mNGS was able to detect TB-genome in nearly half (40/82) of the PE samples from TBP subgroup, with 100% specificity. To evaluate the performance of using CNV features of the human genome for MPE prediction, we performed the leave-one-out cross-validation (LOOCV) in the subcohort excluding the 25 non-MPE samples for setting copy number standards, which demonstrated 54.1% sensitivity, 80.8% specificity, 71.7% accuracy, and an AUC of 0.851.ConclusionIn summary, we exploited the value of human and non-human sequencing reads generated from mNGS, which showed promising ability in simultaneously detecting TBP and MPE.

Space–time method for analyzing transient heat conduction in functionally graded materials

In this study, a novel approach is presented for simulating transient heat conduction in functionally graded materials (FGM) with varying material properties only in the z direction. The proposed collocation method employs a meshless localized space–time radial basis function (LSTRBF), which yields a linear sparse matrix system, offering advantages over the global method in terms of scalability and suitability for multidimensional problems. Additionally, a novel shape parameter strategy, known as variable leave-one-out cross-validation (LOOCV), is introduced to enhance accuracy while overcoming conventional LOOCV limitations. Four benchmark numerical examples are utilized to demonstrate the simplicity, efficiency, accuracy, and stability of the method. Overall, the proposed LSTRBF method presents a promising alternative for efficiently and accurately simulating transient heat conduction in FGMs.

Development of a multi-feature-combined model: proof-of-concept with application to local failure prediction of post-SBRT or surgery early-stage NSCLC patients.

To develop a Multi-Feature-Combined (MFC) model for proof-of-concept in predicting local failure (LR) in NSCLC patients after surgery or SBRT using pre-treatment CT images. This MFC model combines handcrafted radiomic features, deep radiomic features, and patient demographic information in an integrated machine learning workflow. The MFC model comprised three key steps. (1) Extraction of 92 handcrafted radiomic features from the GTV segmented on pre-treatment CT images. (2) Extraction of 512 deep radiomic features from pre-trained U-Net encoder. (3) The extracted handcrafted radiomic features, deep radiomic features, along with 4 patient demographic information (i.e., gender, age, tumor volume, and Charlson comorbidity index), were concatenated as a multi-dimensional input to the classifiers for LR prediction. Two NSCLC patient cohorts from our institution were investigated: (1) the surgery cohort includes 83 patients with segmentectomy or wedge resection (7 LR), and (2) the SBRT cohort includes 84 patients with lung SBRT (9 LR). The MFC model was developed and evaluated independently for both cohorts, and was subsequently compared against the prediction models based on only handcrafted radiomic features (R models), patient demographic information (PI models), and deep learning modeling (DL models). ROC with AUC was adopted to evaluate model performance with leave-one-out cross-validation (LOOCV) and 100-fold Monte Carlo random validation (MCRV). The t-test was performed to identify the statistically significant differences. In LOOCV, the AUC range (surgery/SBRT) of the MFC model was 0.858-0.895/0.868-0.913, which was higher than the three other models: 0.356-0.480/0.322-0.650 for PI models, 0.559-0.618/0.639-0.682 for R models, and 0.809/0.843 for DL models. In 100-fold MCRV, the MFC model again showed the highest AUC results (surgery/SBRT): 0.742-0.825/0.888-0.920, which were significantly higher than PI models: 0.464-0.564/0.538-0.628, R models: 0.557-0.652/0.551-0.732, and DL models: 0.702/0.791. We successfully developed an MFC model that combines feature information from multiple sources for proof-of-concept prediction of LR in patients with surgical and SBRT early-stage NSCLC. Initial results suggested that incorporating pre-treatment patient information from multiple sources improves the ability to predict the risk of local failure.