Search

This paper provides evidence on the use of Random Regression Forests (RRF) for optimal lag selection. Using an extended sample of 144 data series, of various data types with different frequencies and sample sizes, we perform optimal lag selection using RRF and compare the results with seven "traditional" information criteria as well as with three other machine learning approaches. We show that the different information criteria produce differing outcomes in terms of optimal lag selection. To quantify performance, we compare the forecast errors on autoregressive models using the optimal lags selected by the criteria and demonstrate that RRF outperforms other approaches. We provide suggestions to researchers as to which approach to use, under different combinations of data type/data frequency and data type/sample size. [ABSTRACT FROM AUTHOR]

Purpose: To develop a model based on whole-liver radiomics features of pre-treatment enhanced MRI for predicting the prognosis of hepatocellular carcinoma (HCC) patients undergoing continued transarterial chemoembolization (TACE) after TACE-resistance. Materials and Methods: Data from 111 TACE-resistant HCC patients between January 2014 and March 2018 were retrospectively collected. At a ratio of 7:3, patients were randomly assigned to developing and validation cohorts. The whole-liver were manually segmented, and the radiomics signature was extracted. The tumor and liver radiomics score (TLrad-score) was calculated. Models were trained by machine learning algorithms and their predictive efficacies were compared. Results: Tumor stage, tumor burden, body mass index, alpha-fetoprotein, and vascular invasion were revealed as independent risk factors for survival. The model trained by Random Forest algorithms based on tumor burden, whole-liver radiomics signature, and clinical features had the highest predictive efficacy, with c-index values of 0.85 and 0.80 and areas under the ROC curve of 0.96 and 0.83 in the developing cohort and validation cohort, respectively. In the high-rad-score group (TLrad-score > − 0.34), the median overall survival (mOS) was significantly shorter than in the low-rad-score group (17 m vs. 37 m, p < 0.001). A shorter mOS was observed in patients with high tumor burden compared to those with low tumor burden (14 m vs. 29 m, p = 0.007). Conclusion: The combined radiomics model from whole-liver signatures may effectively predict survival for HCC patients continuing TACE after TACE refractoriness. The TLrad-score and tumor burden are potential prognostic markers for TACE therapy following TACE-resistance. [ABSTRACT FROM AUTHOR]

Objectives: We aim to develop a deep learning model based on a convolutional neural network (CNN) combined with a classification algorithm (CA) to assist dentists in quickly and accurately diagnosing the stage of periodontitis. Materials and methods: Periapical radiographs (PERs) and clinical data were collected. The CNNs including Alexnet, VGG16, and ResNet18 were trained on PER to establish the PER-CNN models for no periodontal bone loss (PBL) and PBL. The CAs including random forest (RF), support vector machine (SVM), naive Bayes (NB), logistic regression (LR), and k-nearest neighbor (KNN) were added to the PER-CNN model for control, stage I, stage II and stage III/IV periodontitis. Heat map was produced using a gradient-weighted class activation mapping method to visualize the regions of interest of the PER-Alexnet model. Clustering analysis was performed based on the ten PER-CNN scores and the clinical characteristics. Results: The accuracy of the PER-Alexnet and PER-VGG16 models with the higher performance was 0.872 and 0.853, respectively. The accuracy of the PER-Alexnet + RF model with the highest performance for control, stage I, stage II and stage III/IV was 0.968, 0.960, 0.835 and 0.842, respectively. Heat map showed that the regions of interest predicted by the model were periodontitis bone lesions. We found that age and smoking were significantly related to periodontitis based on the PER-Alexnet scores. Conclusion: The PER-Alexnet + RF model has reached high performance for whole-case periodontal diagnosis. The CNN models combined with CA can assist dentists in quickly and accurately diagnosing the stage of periodontitis. [ABSTRACT FROM AUTHOR]

The influence of hydropriming and Light Emitting Diodes (LED) on germination and growth indices, followed by optimizing and validation via artificial intelligence-based models was carried out in this research. White LEDs (W-LEDs) were more effective by yielding the most effective growth indices, such as mean germination time (MGT) (1.11 day), coefficient of variation of germination time (CVt) (20.72%), mean germination rate (MR) (0.81 day−1), uncertainty (U) (0.40 bit), and synchronization (Z values) (0.79); the optimum MGT (1.09 day), CVt (15.97%), MR (0.77 day−1), U (0.32 bit), and Z (0.55) values were found after 2 h of hydropriming, which was responsible for all efficient growth indicators. W-LEDs with 1 h hydropriming proved to be the ideal LED and hydropriming combination. Results on growth indices for in vitro seedlings were completely different from those on germination indices, and the most desirable germination indices were linked to red LEDs (R-LEDs). Whereas 4 h hydropriming was most effective for the post-germination process. Pareto charts, normal plots, contour plots, and surface plots were created to optimize the input variables. Finally, the data were predicted using Arificial Neural Network (ANN) inspired multilayer perceptron (MLP) and machine learning-based random forest (RF) algorithms. For both models, plant height was correlated with maximum R2 values. Whereas, all output variables had relatively low mean absolute error (MAE), mean square error (MSE), and mean absolute percentage error (MAPE) scores, indicating that both models performed well. The results of this investigation disclosed a link between certain LEDs and hydropriming treatment for in vitro germination indices and plant growth. Graphical presentation of actual and predicted values for germination indices in chickpea [ABSTRACT FROM AUTHOR]

This study examines the effectiveness of various quantile regression (QR) and machine learning (ML) methodologies developed for analyzing the relationship between meteorological parameters and daily reference evapotranspiration (ETref) across diverse climates in Iran spanning from 1987 to 2022. The analyzed models include D-vine copula-based quantile regression (DVQR), multivariate linear quantile regression (MLQR), Bayesian model averaging quantile regression (BMAQR), as well as machine learning algorithms such as extreme learning machine (ELM), random forest (RF), M5 model Tree (M5Tree), least squares support vector regression algorithm (LSSVR), and extreme gradient boosting (XGBoost). Additionally, empirical equations (EEs) such as Baier and Robertson (BARO), Jensen and Haise (JEHA), and Penman (PENM) models were considered. While the EEs demonstrated acceptable performance, the QR and ML models exhibited superior accuracy. Among these, the MLQR model displayed the highest accuracy compared to DVQR and BMAQR models. Moreover, LSSVR, XGBoost, and M5Tree models outperformed ELM and RF models. Notably, LSSVR, XGBoost, and MLQR models exhibited comparable performance (R2 and NSE > 0.92, MBE and RMSE < 0.5, and SI > 0.05) to M5Tree and BMAQR models across all climates. Importantly, these models significantly outperformed EEs, DVQR, ELM, and RF models in all climates. In conclusion, high-dimensional QR and ML models are recommended as promising alternatives for accurately estimating daily ETref in diverse global climate conditions. [ABSTRACT FROM AUTHOR]

As a popular nutraceutical, collagen powder is the target of adulteration like other high-value food products, therefore detection of adulteration in collagen powder is of great practical significance to ensure market order and the health of the people. To achieve accurate detection of adulterants in collagen powder, a modeling method of fluorescence hyperspectral technology combined with machine learning algorithm was proposed. This study employed various preprocessing methods for denoising and spectral correction to enhance the effective spectral data. Principal component analysis was used to visualize the spectral data and initially revealed the spatial distribution of adulterated collagen powder. Genetic algorithm-k nearest neighbor, particle swarm optimization-support vector machine, and gradient boosting decision tree were used to construct classification models to identify adulterated collagen powder, with the best 2-class discriminant model accuracy, 4-class discriminant model accuracy, and 5-class discriminant model accuracy reaching 99%, 94%, and 98%, respectively. In the quantitative detection models of adulteration level, the random forest model performed best with correlation coefficient (R2) of 0.95 to 0.99. These results suggested that fluorescence hyperspectral technology combined with machine learning algorithm can be effectively used to detect adulterated collagen powder. [ABSTRACT FROM AUTHOR]

The proliferation of wireless networks as a primary data transmission channel has brought about a surge in data volume but also raised security threats and privacy concerns. Intrusion Detection Systems (IDS) have proven effective in safeguarding data transmission, yet they struggle to detect minor or rare attacks that mimic normal traffic. To address this challenge, we propose a novel two-layer classification model integrating K-nearest neighbor (KNN) and support vector machines (SVMs). In the first layer, KNN classifies data into Normal, Major, and Minor Attack categories, while the second layer further distinguishes Major Attacks into DoS or Probe and Minor Attacks into U2R or R2. Our framework incorporates Common Correlated Feature Selection (CCFS) to optimize feature discrimination, partitioning training data into three groups. Furthermore, we explore data preprocessing techniques to enhance data interpretation and maintain statistical normalcy in traffic connection features. Our experimental analysis utilizes the NSL-KDD dataset, demonstrating that our approach significantly improves detection rates, particularly for Minor Attacks, achieving a remarkable 92.33% detection rate for U2R and 91.33% for R2L attacks. This represents a substantial 10% enhancement over existing methods, outperforming Multi-Layer Perceptron (MLP) by 13.23%, Random Forest (RF) by 10.11%, J48- Decision Tree (DT) by 9.23%, and Naïve Bayes (NB) by 9.42% and 8.17% in terms of detection rates. These results underscore the effectiveness of our approach in enhancing intrusion detection performance. [ABSTRACT FROM AUTHOR]

Learning from positive and unlabeled data, or PU learning, is the setting in which a binary classifier can only train from positive and unlabeled instances, the latter containing both positive as well as negative instances. Many PU applications, e.g., fraud detection, are also characterized by class imbalance, which creates a challenging setting. Not only are fewer minority class examples compared to the case where all labels are known, there is also only a small fraction of unlabeled observations that would actually be positive. Despite the relevance of the topic, only a few studies have considered a class imbalance setting in PU learning. In this paper, we propose a novel technique that can directly handle imbalanced PU data, named the PU Hellinger Decision Tree (PU-HDT). Our technique exploits the class prior to estimate the counts of positives and negatives in every node in the tree. Moreover, the Hellinger distance is used instead of more conventional splitting criteria because it has been shown to be class-imbalance insensitive. This simple yet effective adaptation allows PU-HDT to perform well in highly imbalanced PU data sets. We also introduce PU Stratified Hellinger Random Forest (PU-SHRF), which uses PU-HDT as its base learner and integrates a stratified bootstrap sampling. Our empirical analysis shows that PU-SHRF substantially outperforms state-of-the-art PU learning methods for imbalanced data sets in most experimental settings. [ABSTRACT FROM AUTHOR]

Levels of a predictor variable that are absent when a classification tree is grown can not be subject to an explicit splitting rule. This is an issue if these absent levels are present in a new observation for prediction. To date, there remains no satisfactory solution for absent levels in random forest models. Unlike missing data, absent levels are fully observed and known. Ordinal encoding of predictors allows absent levels to be integrated and used for prediction. Using a case study on source attribution of Campylobacter species using whole genome sequencing (WGS) data as predictors, we examine how target-agnostic versus target-based encoding of predictor variables with absent levels affects the accuracy of random forest models. We show that a target-based encoding approach using class probabilities, with absent levels designated the highest rank, is systematically biased, and that this bias is resolved by encoding absent levels according to the a priori hypothesis of equal class probability. We present a novel method of ordinal encoding predictors via principal coordinates analysis (PCO) which capitalizes on the similarity between pairs of predictor levels. Absent levels are encoded according to their similarity to each of the other levels in the training data. We show that the PCO-encoding method performs at least as well as the target-based approach and is not biased. [ABSTRACT FROM AUTHOR]

This study presents a pioneering machine learning approach to continuously model fracture intensity in hydrocarbon reservoirs using solely conventional well logs and mud loss data. While machine learning has previously been applied to predict discrete fracture properties, this is among the first attempts to leverage well logs for continuous fracture intensity modeling leveraging advanced ensemble techniques. A multi-level stacked ensemble methodology systematically combines the strengths of diverse algorithms like gradient boosting, random forest and XGBoost through a tiered approach, enhancing predictive performance beyond individual models. Nine base machine learning algorithms generate initial fracture intensity predictions which are combined through linear regression meta-models and further stacked using ridge regression into an integrated super-learner model. This approach achieves significant improvements over individual base models, with the super-learner attaining a mean absolute error of 0.083 and R^2 of 0.980 on test data. By quantifying the crucial fracture intensity parameter continuously as a function of depth, this data-driven methodology enables more accurate reservoir characterization compared to traditional methods. The ability to forecast fracture intensity solely from conventional well logs opens new opportunities for rapid, low-cost quantification of this parameter along new wells without requiring advanced logging tools. When incorporated into reservoir simulators, these machine learning fracture intensity models can help optimize production strategies and recovery management. This systematic stacked ensemble framework advances continuous fracture intensity modeling exclusively from well logs, overcoming limitations of prior techniques. Novel insights gained via rigorous model evaluation deepen the understanding of naturally fractured reservoirs. [ABSTRACT FROM AUTHOR]

Acoustic impedance is the product of the density of a material and the speed at which an acoustic wave travels through it. Understanding this relationship is essential because low acoustic impedance values are closely associated with high porosity, facilitating the accumulation of more hydrocarbons. In this study, we estimate the acoustic impedance based on nine different inputs of seismic attributes in addition to depth and two-way travel time using three supervised machine learning models, namely extra tree regression (ETR), random forest regression, and a multilayer perceptron regression algorithm using the scikit-learn library. Our results show that the R2 of multilayer perceptron regression is 0.85, which is close to what has been reported in recent studies. However, the ETR method outperformed those reported in the literature in terms of the mean absolute error, mean squared error, and root-mean-squared error. The novelty of this study lies in achieving more accurate predictions of acoustic impedance for exploration. [ABSTRACT FROM AUTHOR]

Employing a large dataset (at most, the order of n = 106), this study attempts enhance the literature on the comparison between regression and machine learning-based rent price prediction models by adding new empirical evidence and considering the spatial dependence of the observations. The regression-based approach incorporates the nearest neighbor Gaussian processes (NNGP) model, enabling the application of kriging to large datasets. In contrast, the machine learning-based approach utilizes typical models: extreme gradient boosting (XGBoost), random forest (RF), and deep neural network (DNN). The out-of-sample prediction accuracy of these models was compared using Japanese apartment rent data, with a varying order of sample sizes (i.e., n = 104, 105, 106). The results showed that, as the sample size increased, XGBoost and RF outperformed NNGP with higher out-of-sample prediction accuracy. XGBoost achieved the highest prediction accuracy for all sample sizes and error measures in both logarithmic and real scales and for all price bands if the distribution of rents is similar in training and test data. A comparison of several methods to account for the spatial dependence in RF showed that simply adding spatial coordinates to the explanatory variables may be sufficient. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Orofacial Orthopedics/Fortschritte der Kieferorthopadie is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Conformal prediction has seen many applications in pharmaceutical science, being able to calibrate outputs of machine learning models and producing valid prediction intervals. We here present the open source software CPSign that is a complete implementation of conformal prediction for cheminformatics modeling. CPSign implements inductive and transductive conformal prediction for classification and regression, and probabilistic prediction with the Venn-ABERS methodology. The main chemical representation is signatures but other types of descriptors are also supported. The main modeling methodology is support vector machines (SVMs), but additional modeling methods are supported via an extension mechanism, e.g. DeepLearning4J models. We also describe features for visualizing results from conformal models including calibration and efficiency plots, as well as features to publish predictive models as REST services. We compare CPSign against other common cheminformatics modeling approaches including random forest, and a directed message-passing neural network. The results show that CPSign produces robust predictive performance with comparative predictive efficiency, with superior runtime and lower hardware requirements compared to neural network based models. CPSign has been used in several studies and is in production-use in multiple organizations. The ability to work directly with chemical input files, perform descriptor calculation and modeling with SVM in the conformal prediction framework, with a single software package having a low footprint and fast execution time makes CPSign a convenient and yet flexible package for training, deploying, and predicting on chemical data. CPSign can be downloaded from GitHub at https://github.com/arosbio/cpsign. Scientific contribution CPSign provides a single software that allows users to perform data preprocessing, modeling and make predictions directly on chemical structures, using conformal and probabilistic prediction. Building and evaluating new models can be achieved at a high abstraction level, without sacrificing flexibility and predictive performance—showcased with a method evaluation against contemporary modeling approaches, where CPSign performs on par with a state-of-the-art deep learning based model. [ABSTRACT FROM AUTHOR]

This study evaluated the performance of multiple models that used machine learning to anticipate wind speed in the city of Dhaka. The NASA Power website provided the data set for this investigation. The models used for prediction included the decision tree regressor, support vector regressor, random forest, linear regression, neural network and polynomial regression. A hold-out check and k-fold cross-validation were used to assess how well these models performed. With the highest R2 scores and lowest RMSEs on both the validation and test sets, the results demonstrated that the polynomial regression model performed the best. With the lowest R2 scores and largest RMSEs on both sets, the decision tree model scored the poorest. High R2 scores and low RMSEs were achieved by the random forest model, which had comparable performance to the polynomial regression model but required a longer computation time. In addition, the neural network model demonstrated commendable predictive accuracy, yielding an R2 score of 0.67 and a low RMSE of 0.57. However, its application is contingent on the availability of substantial computational resources, given its extensive computation time of 457.93 s. The study concludes by highlighting the efficacy of the Polynomial Regression model as the optimal choice for wind speed prediction in Dhaka, offering a balance between superior performance and efficient computation. This insight provides valuable guidance for practitioners and researchers seeking effective models for similar applications. [ABSTRACT FROM AUTHOR]

Machine learning is frequently used in various geotechnical applications nowadays. This study presents a statistics and machine learning model for settlement prediction of helical piles that relates compressive service load and soil parameters as a group with the pile parameters. Machine learning algorithms such as Decision Trees, Random Forests, AdaBoost, and Artificial Neural Networks (ANN) were used to develop the predictive models. The models were validated using cross-validation techniques and tested on an independent dataset to assess their accuracy and generalizability. Numerical investigation is used here to supplement the field data by simulating various soil conditions and pile geometries that have not been tested in the field. This study compiled numerical results of 3600 models. As the models are well-calibrated and validated, the data from these models can be reasonably assumed to simulate the ground situation. At the end of this study, a comparative analysis of statistic learning and machine learning (ML) was done using the field axial load tests database and numerical investigation on helical piles. It is observed that ML models like Decision Trees and Random Forests provided the better model with R-squared values of 0.92 and 0.96, respectively, for large diameters. The authors believe this study will permit engineers and state agencies to understand this prediction model's efficacy better, resulting in a more resilient approach to designing large-diameter helical piles for the compressive load. [ABSTRACT FROM AUTHOR]

Random forests (RF) have been widely used to predict spatial variables. Several studies have shown that spatial cross-validation (CV) methods consistently cause RF to yield larger prediction errors compared to standard CV methods. This study examined the impact of species characteristics and data features on the performance of the standard RF and spatial CV approaches for predicting species abundance distribution. It compared the standard 5-fold CV, design-based validation, and three different spatial CV methods, such as spatial buffering, environmental blocking, and spatial blocking. Validation samples were randomly selected for design-based validation without replacement. We evaluated their predictive performance (accuracy and discrimination metrics) using artificial species abundance data generated by a linear function of a constant term ( β 0 ) and a random error term following a zero-mean Gaussian process with a covariance matrix determined by an exponential correlation function. The model was tuned over multiple simulations to consider different mean levels of species abundance, spatial autocorrelation variation, and species detection probability. Here we found that the standard RF had poor predictive performance when spatial autocorrelation was high and the species probability of detection was low. Design-based validation and standard K-fold CV were found to be the most effective strategies for evaluating RF performance compared to spatial CV methods, even in the presence of high spatial autocorrelation and imperfect detection for random samples. For weakly or moderately clustered samples, they yielded good modelling efficiency but overestimated RF's predictive power, while they overestimated modelling efficiency, predictive power, and accuracy for strongly clustered samples with high spatial autocorrelation. Globally, the checkerboard pattern in the allocation of blocks to folds in blocked spatial CV was found to be the most effective CV approach for clustered samples, whatever the degree of clustering, spatial autocorrelation, or species abundance class. The checkerboard pattern in spatial CV was found to be the best method for random or systematic samples with spatial autocorrelation, but less effective than non-spatial CV approaches. Failing to take data features into account when validating models can lead to unrealistic predictions of species abundance and related parameters and, therefore, incorrect interpretations of patterns and conclusions. Further research should explore the benefits of using blocked spatial K-fold CV with checkerboard assignment of blocks to folds for clustered samples with high spatial autocorrelation. [ABSTRACT FROM AUTHOR]

Agriculture in Maharashtra has immense importance in India, acting as the back-bone of the economy and a primary livelihood source for a significant population. Being the third largest state in India, Maharashtra has a high scale crop production in the country which also has an important impact on the economy. Initially the study focus on developing predictive models that guide farmers in selecting suitable crops for the divisions in the state of Maharashtra. This study presents a Crop Recommendation System (CRS) designed to support Maharashtra's agricultural sector by utilizing a comprehensive dataset from 2001 to 2022 provided by the India Meteorological Department. This study helps in improvising technical efficiency and productivity of the farmers. Harvesting crops in optimal condition can help to produce efficient harvest hence the research concentrates on providing best crop recommendation system (CRS) with the help of Machine Learning and Deep Learning techniques. The data, enhanced for accuracy using expectation-maximization optimization, underpins predictive models that guide crop selection. EM contributes to a more robust and reliable dataset for subsequent analyses and modeling by iterative estimating and updating missing values based on probabilistic expectations. Key findings show that the Random Forest algorithm excels in predicting suitable crops with 92% accuracy. Further precision is achieved through a Long Short-Term Memory network forecasting weather patterns three months ahead, accommodating temporal data variations. Subsequently, the proposed system leverages these forecasts to recommend five ideal crops per division within Maharashtra, aiding farmers' decision-making and adapting to regional climatic conditions. A supplementary crop calendar offers monthly district-specific planting guidance. An intuitive Graphical User Interface delivers this information effectively, ensuring practical and informed agricultural choices across the state. In essence, the study provides an innovative tool for enhancing economic stability and sustenance in Maharashtra through technology-driven agriculture recommendations aligned with future weather expectations. [ABSTRACT FROM AUTHOR]

The present study investigated whether life engagement and happiness can be predicted from gaming motives and primary emotional traits. Two machine learning algorithms (random forest model and one-dimensional convolutional neural network) were applied using a dataset from before the COVID-19 pandemic as the training dataset. The algorithms derived were then applied to test if they would be useful in predicting life engagement and happiness from gaming motives and primary emotional systems on a dataset collected during the pandemic. The best prediction values were observed for happiness with ρ = 0.758 with explained variance of R2 = 0.575 when applying the best performing algorithm derived from the pre-COVID dataset to the COVID dataset. Hence, this shows that the derived algorithm based on the pre-pandemic data set, successfully predicted happiness (and life engagement) from the same set of variables during the pandemic. Overall, this study shows the feasibility of applying machine learning algorithms to predict life engagement and happiness from gaming motives and primary emotional systems. [ABSTRACT FROM AUTHOR]

As urban development accelerates and natural disasters occur more frequently, the urgency of developing effective emergency shelter planning strategies intensifies. The shelter location selection method under the traditional multi-criteria decision-making framework suffers from issues such as strong subjectivity and insufficient data support. Artificial intelligence offers a robust data-driven approach for site selection; however, many methods neglect the spatial relationships of site selection targets within geographical space. This paper introduces an emergency shelter site selection model that combines a variational graph autoencoder (VGAE) with a random forest (RF), namely VGAE-RF. In the constructed urban spatial topological graph, based on network geographic information, this model captures both the latent features of geographic unit coupling and integrates explicit and latent features to forecast the likelihood of emergency shelters in the construction area. This study takes Beijing, China, as the experimental area and evaluates the reliability of different model methods using a confusion matrix, Receiver Operating Characteristic (ROC) curve, and Imbalance Index of spatial distribution as evaluation indicators. The experimental results indicate that the proposed VGAE-RF model method, which considers spatial semantic associations, displays the best reliability. [ABSTRACT FROM AUTHOR]

Global patterns of leaf nitrogen (N) and phosphorus (P) stoichiometry have been interpreted as reflecting phenotypic plasticity in response to the environment, or as an overriding effect of the distribution of species growing in their biogeochemical niches. Here, we balance these contrasting views. We compile a global dataset of 36,413 paired observations of leaf N and P concentrations, taxonomy and 45 environmental covariates, covering 7,549 sites and 3,700 species, to investigate how species identity and environmental variables control variations in mass-based leaf N and P concentrations, and the N:P ratio. We find within-species variation contributes around half of the total variation, with 29%, 31%, and 22% of leaf N, P, and N:P variation, respectively, explained by environmental variables. Within-species plasticity along environmental gradients varies across species and is highest for leaf N:P and lowest for leaf N. We identified effects of environmental variables on within-species variation using random forest models, whereas effects were largely missed by widely used linear mixed-effect models. Our analysis demonstrates a substantial influence of the environment in driving plastic responses of leaf N, P, and N:P within species, which challenges reports of a fixed biogeochemical niche and the overriding importance of species distributions in shaping global patterns of leaf N and P. To assess drivers underpinning biogeographic patterns of leaf nitrogen and phosphorus stoichiometry, this study used a global dataset of leaf nitrogen and phosphorus concentrations of 3,625 species from 7,549 sites. They found strong within-species variation related to environmental gradients, challenging the idea of a fixed biogeochemical niche. [ABSTRACT FROM AUTHOR]

Copyright of Fire Ecology is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Background: Non-steroidal anti-inflammatory drugs (NSAIDs) are considered first-line medications for acute migraine attacks. However, the response exhibits considerable variability among individuals. Thus, this study aimed to explore a machine learning model based on the percentage of amplitude oscillations (PerAF) and gray matter volume (GMV) to predict the response to NSAIDs in migraine treatment. Methods: Propensity score matching was adopted to match patients having migraine with response and nonresponse to NSAIDs, ensuring consistency in clinical characteristics and migraine-related features. Multimodal magnetic resonance imaging was employed to extract PerAF and GMV, followed by feature selection using the least absolute shrinkage and selection operator regression and recursive feature elimination algorithms. Multiple predictive models were constructed and the final model with the smallest predictive residuals was chosen. The model performance was evaluated using the area under the receiver operating characteristic (ROCAUC) curve, area under the precision-recall curve (PRAUC), balance accuracy (BACC), sensitivity, F1 score, positive predictive value (PPV), and negative predictive value (NPV). External validation was performed using a public database. Then, correlation analysis was performed between the neuroimaging predictors and clinical features in migraine. Results: One hundred eighteen patients with migraine (59 responders and 59 non-responders) were enrolled. Six features (PerAF of left insula and left transverse temporal gyrus; and GMV of right superior frontal gyrus, left postcentral gyrus, right postcentral gyrus, and left precuneus) were observed. The random forest model with the lowest predictive residuals was selected and model metrics (ROCAUC, PRAUC, BACC, sensitivity, F1 score, PPV, and NPV) in the training and testing groups were 0.982, 0.983, 0.927, 0.976, 0.930, 0.889, and 0.973; and 0.711, 0.648, 0.639, 0.667,0.649, 0.632, and 0.647, respectively. The model metrics of external validation were 0.631, 0.651, 0.611, 0.808, 0.656, 0.553, and 0.706. Additionally, a significant positive correlation was found between the GMV of the left precuneus and attack time in non-responders. Conclusions: Our findings suggest the potential of multimodal neuroimaging features in predicting the efficacy of NSAIDs in migraine treatment and provide novel insights into the neural mechanisms underlying migraine and its optimized treatment strategy. [ABSTRACT FROM AUTHOR]

Neurodevelopmental disorders (NDD) in offspring are associated with a complex combination of pre-and postnatal factors. This study uses machine learning and population data to evaluate the association between prepregnancy or perinatal risk factors and the NDD of offspring. Population-based retrospective cohort data were obtained from Korea National Health Insurance Service claims data for 209,424 singleton offspring and their mothers who gave birth for the first time in 2007. The dependent variables were motor development disorder (MDD), cognitive development disorder (CDD) and combined overall neurodevelopmental disorder (NDD) from offspring. Seventeen independent variables from 2002 to 2007 were included. Random forest variable importance and Shapley Additive Explanation (SHAP) values were calculated to analyze the directions of its associations with the predictors. The random forest with oversampling registered much higher areas under the receiver-operating-characteristic curves than the logistic regression of interaction and non-linearity terms, 79% versus 50% (MDD), 82% versus 52% (CDD) and 74% versus 50% (NDD). Based on random forest variable importance, low socioeconomic status and age at birth were highly ranked. In SHAP values, there was a positive association between NDD and pre- or perinatal outcomes, especially, fetal male sex with growth restriction associated the development of NDD in offspring. [ABSTRACT FROM AUTHOR]

Predicting postpartum hemorrhage (PPH) before delivery is crucial for enhancing patient outcomes, enabling timely transfer and implementation of prophylactic therapies. We attempted to utilize machine learning (ML) using basic pre-labor clinical data and laboratory measurements to predict postpartum Hemoglobin (Hb) in non-complicated singleton pregnancies. The local databases of two academic care centers on patient delivery were incorporated into the current study. Patients with preexisting coagulopathy, traumatic cases, and allogenic blood transfusion were excluded from all analyses. The association of pre-delivery variables with 24-h post-delivery hemoglobin level was evaluated using feature selection with Elastic Net regression and Random Forest algorithms. A suite of ML algorithms was employed to predict post-delivery Hb levels. Out of 2051 pregnant women, 1974 were included in the final analysis. After data pre-processing and redundant variable removal, the top predictors selected via feature selection for predicting post-delivery Hb were parity (B: 0.09 [0.05–0.12]), gestational age, pre-delivery hemoglobin (B:0.83 [0.80–0.85]) and fibrinogen levels (B:0.01 [0.01–0.01]), and pre-labor platelet count (B*1000: 0.77 [0.30–1.23]). Among the trained algorithms, artificial neural network provided the most accurate model (Root mean squared error: 0.62), which was subsequently deployed as a web-based calculator: https://predictivecalculators.shinyapps.io/ANN-HB. The current study shows that ML models could be utilized as accurate predictors of indirect measures of PPH and can be readily incorporated into healthcare systems. Further studies with heterogenous population-based samples may further improve the generalizability of these models. [ABSTRACT FROM AUTHOR]

Leptospirosis is a global disease that impacts people worldwide, particularly in humid and tropical regions, and is associated with significant socio-economic deficiencies. Its symptoms are often confused with other syndromes, which can compromise clinical diagnosis and the failure to carry out specific laboratory tests. In this respect, this paper presents a study of three algorithms (Decision Tree, Random Forest and Adaboost) for predicting the outcome (cure or death) of individuals with leptospirosis. Using the records contained in the government National System of Aggressions and Notification (SINAN, in portuguese) from 2007 to 2017, for the state of Pará, Brazil, where the temporal attributes of health care, symptoms (headache, vomiting, jaundice, calf pain) and clinical evolution (renal failure and respiratory changes) were used. In the performance evaluation of the selected models, it was observed that the Random Forest exhibited an accuracy of 90.81% for the training dataset, considering the attributes of experiment 8, and the Decision Tree presented an accuracy of 74.29 for the validation database. So, this result considers the best attributes pointed out by experiment 10: time first symptoms medical attention, time first symptoms ELISA sample collection, medical attention hospital admission time, headache, calf pain, vomiting, jaundice, renal insufficiency, and respiratory alterations. The contribution of this article is the confirmation that artificial intelligence, using the Decision Tree model algorithm, depicting the best choice as the final model to be used in future data for the prediction of human leptospirosis cases, helping in the diagnosis and course of the disease, aiming to avoid the evolution to death. [ABSTRACT FROM AUTHOR]

The current prostate cancer (PCa) screen test, prostate-specific antigen (PSA), has a high sensitivity for PCa but low specificity for high-risk, clinically significant PCa (csPCa), resulting in overdiagnosis and overtreatment of non-csPCa. Early identification of csPCa while avoiding unnecessary biopsies in men with non-csPCa is challenging. We built an optimized machine learning platform (ClarityDX) and showed its utility in generating models predicting csPCa. Integrating the ClarityDX platform with blood-based biomarkers for clinically significant PCa and clinical biomarker data from a 3448-patient cohort, we developed a test to stratify patients' risk of csPCa; called ClarityDX Prostate. When predicting high risk cancer in the validation cohort, ClarityDX Prostate showed 95% sensitivity, 35% specificity, 54% positive predictive value, and 91% negative predictive value, at a ≥ 25% threshold. Using ClarityDX Prostate at this threshold could avoid up to 35% of unnecessary prostate biopsies. ClarityDX Prostate showed higher accuracy for predicting the risk of csPCa than PSA alone and the tested model-based risk calculators. Using this test as a reflex test in men with elevated PSA levels may help patients and their healthcare providers decide if a prostate biopsy is necessary. [ABSTRACT FROM AUTHOR]

The classification of imbalanced datasets is a prominent task in text mining and machine learning. The number of samples in each class is not uniformly distributed; one class contains a large number of samples while the other has a small number. Overfitting of the model occurs as a result of imbalanced datasets, resulting in poor performance. In this study, we compare different oversampling techniques like synthetic minority oversampling technique (SMOTE), support vector machine SMOTE (SVM-SMOTE), Border-line SMOTE, K-means SMOTE, and adaptive synthetic (ADASYN) oversampling to address the issue of imbalanced datasets and enhance the performance of machine learning models. Preprocessing significantly enhances the quality of input data by reducing noise, redundant data, and unnecessary data. This enables the machines to identify crucial patterns that facilitate the extraction of significant and pertinent information from the preprocessed data. This study preprocesses the data using various top-level preprocessing steps. Furthermore, two imbalanced Twitter datasets are used to compare the performance of oversampling techniques with six machine learning models including random forest (RF), SVM, K-nearest neighbor (KNN), AdaBoost (ADA), logistic regression (LR), and decision tree (DT). In addition, the bag of words (BoW) and term frequency and inverse document frequency (TF-IDF) features extraction approaches are used to extract features from the tweets. The experiments indicate that SMOTE and ADASYN perform much better than other techniques thus providing higher accuracy. Additionally, overall results show that SVM with 'linear' kernel tends to attain the highest accuracy and recall score of 99.67% and 1.00% on ADASYN oversampled datasets and 99.57% accuracy on SMOTE oversampled dataset with TF-IDF features. The SVM model using 10-fold cross-validation experiments achieved 97.40 mean accuracy with a 0.008 standard deviation. Our approach achieved 2.62% greater accuracy as compared to other current methods. [ABSTRACT FROM AUTHOR]

Skin lesion classification plays a crucial role in the early detection and diagnosis of various skin conditions. Recent advances in computer-aided diagnostic techniques have been instrumental in timely intervention, thereby improving patient outcomes, particularly in rural communities lacking specialized expertise. Despite the widespread adoption of convolutional neural networks (CNNs) in skin disease detection, their effectiveness has been hindered by the limited size and data imbalance of publicly accessible skin lesion datasets. In this context, a two-step hierarchical binary classification approach is proposed utilizing hybrid machine and deep learning (DL) techniques. Experiments conducted on the International Skin Imaging Collaboration (ISIC 2017) dataset demonstrate the effectiveness of the hierarchical approach in handling large class imbalances. Specifically, employing DenseNet121 (DNET) as a feature extractor and random forest (RF) as a classifier yielded the most promising results, achieving a balanced multiclass accuracy (BMA) of 91.07% compared to the pure deep-learning model (end-to-end DNET) with a BMA of 88.66%. The RF ensemble exhibited significantly greater efficiency than other machine-learning classifiers in aiding DL to address the challenge of learning with limited data. Furthermore, the implemented predictive hybrid hierarchical model demonstrated enhanced performance while significantly reducing computational time, indicating its potential efficiency in real-world applications for the classification of skin lesions. [ABSTRACT FROM AUTHOR]

Identifying the key determinants of heavy metal(loid) accumulation in rice and quantifying their contributions are critical for precise prediction of heavy metal(loid) concentrations in rice and the formulation of effective pollution control strategies. The accumulation of heavy metal(loid)s in rice can be influenced by both natural and anthropogenic factors, which may interact with each other. However, distinguishing the independent roles (main effects) from interactive effects and quantifying their impacts separately pose challenges. To address this knowledge gap, we employed TreeExplainer-based SHAP and random forest algorithms in this study to quantitatively estimate the primary influencing factors and their main and interactive effects on heavy metal(loid)s in rice. Our findings reveal that soil cadmium (SCd) and rice cultivation time (C_TIME) were the primary contributors to rice cadmium (RCd) and rice arsenic (RAs), respectively. Soil lead (SPb) and sampling distances from roads significantly contributed to rice lead (RPb). Additionally, we identified significant interactive effects of SCd and C_TIME, C_TIME and RCd, and RCd and rice variety on RCd, RAs, and RPb, respectively, emphasizing their significance. These insights are pivotal in improving the accuracy of heavy metal(loid) concentration predictions in rice and offering theoretical guidance for the formulation of pollution control measures. [ABSTRACT FROM AUTHOR]

Nano-phytoremediation is a novel green technique to remove toxic pollutants from the environment. In vitro regenerated Ceratophyllum demersum (L.) plants were exposed to different concentrations of chromium (Cr) and exposure times in the presence of titania nanoparticles (TiO2NPs). Response surface methodology was used for multiple statistical analyses like regression analysis and optimizing plots. The supplementation of NPs significantly impacted Cr in water and Cr removal (%), whereas NP × exposure time (T) statistically regulated all output parameters. The Firefly metaheuristic algorithm and the random forest (Firefly-RF) machine learning algorithms were coalesced to optimize hyperparameters, aiming to achieve the highest level of accuracy in predicted models. The R2 scores were recorded as 0.956 for Cr in water, 0.987 for Cr in the plant, 0.992 for bioconcentration factor (BCF), and 0.957 for Cr removal through the Firefly-RF model. The findings illustrated superior prediction performance from the random forest models when compared to the response surface methodology. The conclusion is drawn that metal-based nanoparticles (NPs) can effectively be utilized for nano-phytoremediation of heavy metals. This study has uncovered a promising outlook for the utilization of nanoparticles in nano-phytoremediation. This study is expected to pave the way for future research on the topic, facilitating further exploration of various nanoparticles and a thorough evaluation of their potential in aquatic ecosystems. [ABSTRACT FROM AUTHOR]

Tree-based machine learning algorithms, such as random forest, have emerged as effective tools for estimating fine particulate matter (PM2.5) from satellite observations. However, they typically have unchanged model structures and configurations over time and space, and thus may not fully capture the spatiotemporal variations in the relationship between PM2.5 and predictors, resulting in limited accuracy. Here, we propose geographically and temporally weighted tree-based models (GTW-Tree) for remote sensing of surface PM2.5. Unlike traditional tree-based models, GTW-Tree models vary by time and space to simulate the variability in PM2.5 estimation, and they can output variable importance for every location for the deeper understanding of PM2.5 determinants. Experiments in China demonstrate that GTW-Tree models significantly outperform the conventional tree-based models with predictive error reduced by >21%. The GTW-Tree-derived time-location-specific variable importance reveals spatiotemporally varying impacts of predictors on PM2.5. Aerosol optical depth (AOD) contributes largely to PM2.5 estimation, particularly in central China. The proposed models are valuable for spatiotemporal modeling and interpretation of PM2.5 and other various fields of environmental remote sensing. [ABSTRACT FROM AUTHOR]

This study introduces a novel Hybrid Ensemble Machine-Learning (HEML) algorithm to merge long-term satellite-based reanalysis precipitation products (SRPPs), enabling the estimation of super drought events in the Lake Victoria Basin (LVB) during the period of 1984 to 2019. This study considers three widely used Machine learning (ML) models, including RF (Random Forest), GBM (Gradient Boosting Machine), and KNN (k-nearest Neighbors), for the emerging HEML approach. The three SRPPs, including CHIRPS (Climate Hazards Group Infra-Red Precipitation with Station), ERA5-Land, and PERSIANN-CDR (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Network-Climate Data Record), were used to merge for developing new precipitation estimates from HEML model. Additionally, classification and regression models were employed as base learners in developing this algorithm. The newly developed HEML datasets were compared with other ML and SRPP products for super-drought monitoring. The Standardized precipitation evapotranspiration index (SPEI) was used to estimate super drought characteristics, including Drought frequency (DF), Drought Duration (DD), and Drought Intensity (DI) from machine learning and SRPPs products in LVB and compared with RG observation. The results revealed that the HEML algorithm shows excellent performance (CC = 0.93) compared to the single ML merging method and SRPPs against observation. Furthermore, the HEML merging product adeptly captures the spatiotemporal patterns of super drought characteristics during both training (1984–2009) and testing (2010–2019) periods. This research offers crucial insights for near-real-time drought monitoring, water resource management, and informed policy decisions. [ABSTRACT FROM AUTHOR]

Anticipating human decisions while performing complex tasks remains a formidable challenge. This study proposes a multimodal machine-learning approach that leverages image features and electroencephalography (EEG) data to predict human response correctness in a demanding visual searching task. Notably, we extract a novel set of image features pertaining to object relationships using the Segment Anything Model (SAM), which enhances prediction accuracy compared to traditional features. Additionally, our approach effectively utilizes a combination of EEG signals and image features to streamline the feature set required for the Random Forest Classifier (RFC) while maintaining high accuracy. The findings of this research hold substantial potential for developing advanced fault alert systems, particularly in critical decision-making environments such as the medical and defence sectors. [ABSTRACT FROM AUTHOR]

This study aims to compare the performance of ensembles according to their inherent diversity in the context of landslide susceptibility assessment. Heterogeneous and homogeneous ensemble types can be distinguished; four ensembles of each approach were implemented in the Djebahia region. The heterogeneous ensembles include stacking (ST), voting (VO), weighting (WE), and a new approach in landslide assessment called meta-dynamic ensemble selection (DES), while the homogeneous ensembles include AdaBoost (ADA), bagging (BG), random forest (RF), and random subspace (RSS). To ensure a consistent comparison, each ensemble was implemented using individual base learners. The heterogeneous ensembles were generated by combining eight different machine learning algorithms, while the homogeneous ensembles only used a single base learner, with diversity achieved through resampling the training dataset. The spatial dataset used in this study consisted of 115 landslide events and 12 conditioning factors, which were randomly divided into training and testing datasets. The models were evaluated through various aspects, including receiver operating characteristic (ROC) curves, root mean squared error (RMSE), landslide density distribution (LDD), threshold-dependent metrics (Kappa index, accuracy, and recall scores), and a global visual representation using the Taylor diagram. Additionally, a sensitivity analysis (SA) was conducted for the best performing models to assess the importance of the factors and the resilience of the ensembles. The results revealed that homogeneous ensembles outperformed heterogeneous ensembles in terms of AUC and threshold-dependent metrics, with AUC ranging from 0.962 to 0.971 for the test dataset. ADA was the best performing model for these metrics and the least in terms of RMSE (0.366). However, the heterogeneous ensemble ST provided a finer RMSE (0.272), and DES showed the best LDD, indicating a stronger potential to generalize the phenomenon. The Taylor diagram was consistent with the other results, indicating that ST was the best performing model, followed by RSS. The SA demonstrated that RSS was the most robust (mean AUC variation of − 0.022) and ADA was the least robust (mean AUC variation of − 0.038). [ABSTRACT FROM AUTHOR]

In Russian, a specialized reflexive possessive pronoun must be used to indicate coreference between the possessor and the subject. However, it has been observed that in the first and second person plural, the reflexivization rule is often violated, and a non-reflexive possessive is used for subject coreference. Existing studies do not distinguish between singular formal and plural informal readings in the second person, assuming no difference between the two contexts. Furthermore, there is limited knowledge regarding the use of possessives in imperative contexts, as previous research has primarily focused on indicatives. This article aims to investigate the use of possessive pronouns in both formal and informal imperative contexts by addressing two questions. First, how often is a non-reflexive possessive used to indicate coreference with the subject in Russian imperative sentences? Second, what combination of semantic and pragmatic factors allows its use? Apart from the semantic factors that are well documented in the existing literature (referentiality, animacy and alienability of the possessum), we will consider three pragmatic dimensions of politeness that have not been considered so far: form or address (informal t y / formal V y ), type of imperative (speaker- / hearer-oriented) and verbal aspect (perfective / imperfective). The study analyzed the use of possessive pronouns in imperatives in the spoken subcorpus of the Russian National Corpus using generalized linear mixed models and random forests. The findings suggest that the choice of possessive is strongly influenced by the form of address (t y / V y ). When addressing the listener with the informal t y , a reflexive possessive is almost exclusively used. When addressing the listener with the formal V y , both possessives are used equally often. The choice between them depends on several factors, including the referentiality and alienability of the possessum, the type of imperative, and the verbal aspect. The non-reflexive form is preferred when the possessum is referential, the imperative is speaker-oriented, and, to a lesser extent, when the possessum is alienable and the verb is in the perfective aspect. Additionally, some verbs such as 'give' and 'show' strongly prefer a non-reflexive possessive. The results suggest that a reflexive possessive has become the default for subject coreference in Russian second person singular informal contexts. In formal contexts, however, the change is still ongoing, and a non-reflexive possessive is used to express the referential uniqueness of the possessum and to convey negative politeness towards the addressee. [ABSTRACT FROM AUTHOR]

Malnutrition is a frequently underdiagnosed condition leading to increased morbidity, mortality, and healthcare costs. The Mount Sinai Health System (MSHS) deployed a machine learning model (MUST-Plus) to detect malnutrition upon hospital admission. However, in diverse patient groups, a poorly calibrated model may lead to misdiagnosis, exacerbating health care disparities. We explored the model's calibration across different variables and methods to improve calibration. Data from adult patients admitted to five MSHS hospitals from January 1, 2021 - December 31, 2022, were analyzed. We compared MUST-Plus prediction to the registered dietitian's formal assessment. Hierarchical calibration was assessed and compared between the recalibration sample (N = 49,562) of patients admitted between January 1, 2021 - December 31, 2022, and the hold-out sample (N = 17,278) of patients admitted between January 1, 2023 - September 30, 2023. Statistical differences in calibration metrics were tested using bootstrapping with replacement. Before recalibration, the overall model calibration intercept was −1.17 (95% CI: −1.20, −1.14), slope was 1.37 (95% CI: 1.34, 1.40), and Brier score was 0.26 (95% CI: 0.25, 0.26). Both weak and moderate measures of calibration were significantly different between White and Black patients and between male and female patients. Logistic recalibration significantly improved calibration of the model across race and gender in the hold-out sample. The original MUST-Plus model showed significant differences in calibration between White vs. Black patients. It also overestimated malnutrition in females compared to males. Logistic recalibration effectively reduced miscalibration across all patient subgroups. Continual monitoring and timely recalibration can improve model accuracy. [ABSTRACT FROM AUTHOR]

To enhance the real-time monitoring and early-warning capabilities for dust disasters in underground coal mine, this paper presents a novel WGAN-CNN-based prediction approach to predict the dust concentration at underground coal mine working faces. Dust concentration, wind speed, temperature, and methane concentration were collected as the original data due to their nonlinear relationship. The consistency between the generated and original data distributions was verified through PCA dimensionality reduction analysis. The predictive performance of this approach was assessed using five metrics (R2, EVS, MSE, RMSE, and MAE) and compared with three other algorithms (Random Forest Regressor, MLP Regressor, and LinearSVR). The findings indicate that a majority of the generated data falls within the distribution range of the real dataset, exhibiting reduced levels of volatility and dispersion. The R2 values of prediction results are all above 98%, and the MSE values are between 0.0007 and 0.0106. The proposed approach exhibits superior predictive accuracy and robust model generalization capabilities compared to alternative algorithms, thereby enhancing the real-time monitoring and early-warning level of dust disasters in underground coal mine. This will facilitate the realization of advanced prevention and control measures for dust disasters, showcasing a wide range of potential applications. [ABSTRACT FROM AUTHOR]

Tabular data analysis is a critical task in various domains, enabling us to uncover valuable insights from structured datasets. While traditional machine learning methods can be used for feature engineering and dimensionality reduction, they often struggle to capture the intricate relationships and dependencies within real-world datasets. In this paper, we present Multi-representation DeepInsight (MRep-DeepInsight), a novel extension of the DeepInsight method designed to enhance the analysis of tabular data. By generating multiple representations of samples using diverse feature extraction techniques, our approach is able to capture a broader range of features and reveal deeper insights. We demonstrate the effectiveness of MRep-DeepInsight on single-cell datasets, Alzheimer's data, and artificial data, showcasing an improved accuracy over the original DeepInsight approach and machine learning methods like random forest, XGBoost, LightGBM, FT-Transformer and L2-regularized logistic regression. Our results highlight the value of incorporating multiple representations for robust and accurate tabular data analysis. By leveraging the power of diverse representations, MRep-DeepInsight offers a promising new avenue for advancing decision-making and scientific discovery across a wide range of fields. [ABSTRACT FROM AUTHOR]

Identifying a fish's artificial or natural origin is crucial for stock enhancement assessments. In this study, we collected specimens of hatchery (including the early fry, reared, and released juveniles) and wild mangrove snapper Lutjanus argentimaculatus to explore the feasibility of using δ13C and δ15N values in scales as a non-lethal alternative to muscle tissues for discriminating between hatchery and wild origins. The results showed scales generally enriched in δ13C value and relatively consistent in δ15N value compared to muscle. Significant correlations were found between the two tissues for both δ13C and δ15N values (R2 = 0.80–0.88). Based on the isotopic composition of the scale core, individuals with different origins could be effectively discriminated, achieving a classification success rate of 77.6%. The random forest analysis revealed that a δ13C value of −15.8‰ could be a reference for determining the hatchery and wild origins of mangrove snapper. This study provides an isotopic approach for assessing the benefit of stock enrichment in mangrove snapper. It highlights the importance of regular sampling and scales analysis to expand fishery management and conservation practices. [ABSTRACT FROM AUTHOR]

Society's support upon chemicals over the last few decades has led to their increased production, application and discharge into the environment. Wastewater treatment plants (WWTPs) contain a multitude of these chemicals such us; pharmaceutical compounds (PCs). Often, their biodegradability by activated sludge microorganisms is significant for their elimination during wastewater treatment. In this paper the focus is laid on two PCs carbamazepine (CBZ) and diclofenac (DCF) and their main transformation products (TPs). Laboratory degradation tests with these two pharmaceuticals using activated sludge as inoculum under aerobic conditions were performed and microbial metabolites were analyzed by liquid chromatography-mass spectrometry (LC/MS-MS). In two different Mixed liquid Suspended Solids (MLSS) concentrations the biodegradability by activated sludge of CBZ and DCF were evaluated. Also, this article proposes a decision support system to optimize the prediction process of this type of pharmacological compounds. A study and analysis of the techniques of Support Vector Machine, Random Forest, Decision Trees and Multilayer Perceptron Network is carried out to select the most reliable and accurate predictor for the decision system. There are not significant differences in the removal of DCF with 30 mg MLSS/L and 60 mg MLSS/L. DCF was better removed than CBZ in all experiments studied. The TP detected in the samples were mainly 4-OH-DCF for DCF and 10, 11 EPOXICBZ for CBZ. The results show that the best models are obtained with Random Forest and Multilayer Perceptron Network techniques, with a model fit of more than 95% for both carbamazepine and diclofenac metabolites. Obtaining a root means square errors of 0.80 µg/L for the metabolite 4-OH-DCF for DCF with the technique Random Forest and a root means square errors of 1.13 µg/L for the metabolite 10, 11 EPOXICBZ for CBZ with the Multilayer Perceptron Network technique. [ABSTRACT FROM AUTHOR]

Bonica's (Am J Polit Sci 58(2):367–386, 2014) campaign finance-based ideology scores, or CFscores, create an ideological common space that allows researchers to compare a wide variety of actors. Because relatively few citizens donate to candidates, however, the public is not well represented in this common space. This paper addresses that gap. It uses random forest machine learning on data from the 2012 Cooperative Congressional Election Study to impute CFscores for respondents who did not donate to candidates, based on how their policy views compared to those who did. These new scores are robust to differences in issue importance between donors and non-donors, and they outperform other ideological measures in predicting vote choice. The scores are then applied to a substantive exercise. Past research shows that extreme candidates for governor are penalized more by voters than those in lower-profile races. The implied mechanism—that vote choice for governor is more ideologically-driven—can be directly tested with imputed CFscores, since they uniquely allow comparisons between voters and candidates across races. An analysis of voting behavior in 2012 confounds expectations. Ideology appears to factor no more into vote choice for governor than for US House. These novel findings underscore the value of extending CFscores to non-donating survey respondents, and while current efforts are limited by data availability, this study offers encouragement and a roadmap to that end. [ABSTRACT FROM AUTHOR]

High-throughput whole genome sequencing (WGS) is clinically used in finding single nucleotide variants and small indels. Several bioinformatics tools are developed to call short tandem repeats (STRs) copy numbers from WGS data, such as ExpansionHunter denovo, GangSTR and HipSTR. However, expansion disorders are rare and it is hard to find candidate expansions in single patient sequencing data with ~ 800,000 STRs calls. In this paper I describe a snakemake pipeline for genome-wide STRs Annotation and Score (STRAS) using a Random Forest (RF) model to predict pathogenicity. The predictor was validated by benchmark data from Clinvar and PUBMED. True positive rate was 93.8%. True negative rate was 98.0%.Precision was 98.6% and recall rate was 93.8%. F1-score was 0.961. Sensitivity was 93.8% and specificity was 99.6%. These results showed STRAS could be a useful tool for clinical researchers to find STR loci of interest and filter out neutral STRs. STRAS is freely available at https://github.com/fancheyu5/STRAS. [ABSTRACT FROM AUTHOR]

The paper explores the possibilities of using data classification methods when forecasting time series of the geomagnetic Kp-index by machine learning methods. To classify categories of the Kp-index based on the degree of disturbance, linear and logistic regression, random forest, gradient boosting on top of decision trees, and artificial neural networks of various architectures are used. The results of these methods are compared with a trivial inertial forecast (the statistical indicators of which for problems of this type are always high) at horizons from 3 h to 1 day in 3-h increments. The problem of choosing a cross-validation scheme for selecting the model hyperparameters, ways to overcome the imbalance of categories, the relative importance of input features, as well as the dependence of the results on the test sample (beginning of the 25th solar activity cycle) on inclusion in the training sample of data from the 23rd and 24th cycles or only the 24th cycles are studied. Based on the results, conclusions are drawn about the preferred methods for classifying values of the Kp-index based on the level of geomagnetic disturbance. Ways for further research and possible improvement of the classification quality are outlined, including for determining the characteristic hidden states of Earth's magnetosphere as a dynamic system in order to improve the quality of forecasting geomagnetic indices. [ABSTRACT FROM AUTHOR]

In order to study how to use pulmonary functional imaging obtained through 4D-CT fusion for radiotherapy planning, and transform traditional dose volume parameters into functional dose volume parameters, a functional dose volume parameter model that may reduce level 2 and above radiation pneumonia was obtained. 41 pulmonary tumor patients who underwent 4D-CT in our department from 2020 to 2023 were included. MIM Software (MIM 7.0.7; MIM Software Inc., Cleveland, OH, USA) was used to register adjacent phase CT images in the 4D-CT series. The three-dimensional displacement vector of CT pixels was obtained when changing from one respiratory state to another respiratory state, and this three-dimensional vector was quantitatively analyzed. Thus, a color schematic diagram reflecting the degree of changes in lung CT pixels during the breathing process, namely the distribution of ventilation function strength, is obtained. Finally, this diagram is fused with the localization CT image. Select areas with Jacobi > 1.2 as high lung function areas and outline them as fLung. Import the patient's DVH image again, fuse the lung ventilation image with the localization CT image, and obtain the volume of fLung different doses (V60, V55, V50, V45, V40, V35, V30, V25, V20, V15, V10, V5). Analyze the functional dose volume parameters related to the risk of level 2 and above radiation pneumonia using R language and create a predictive model. By using stepwise regression and optimal subset method to screen for independent variables V35, V30, V25, V20, V15, and V10, the prediction formula was obtained as follows: Risk = 0.23656–0.13784 * V35 + 0.37445 * V30-0.38317 * V25 + 0.21341 * V20-0.10209 * V15 + 0.03815 * V10. These six independent variables were analyzed using a column chart, and a calibration curve was drawn using the calibrate function. It was found that the Bias corrected line and the Apparent line were very close to the Ideal line, The consistency between the predicted value and the actual value is very good. By using the ROC function to plot the ROC curve and calculating the area under the curve: 0.8475, 95% CI 0.7237–0.9713, it can also be determined that the accuracy of the model is very high. In addition, we also used Lasso method and random forest method to filter out independent variables with different results, but the calibration curve drawn by the calibration function confirmed poor prediction performance. The function dose volume parameters V35, V30, V25, V20, V15, and V10 obtained through 4D-CT are key factors affecting radiation pneumonia. Establishing a predictive model can provide more accurate lung restriction basis for clinical radiotherapy planning. [ABSTRACT FROM AUTHOR]

The classification of land use and land cover (LULC) from remotely sensed imagery in semi-arid Mediterranean areas is a challenging task due to the fragmentation of the landscape and the diversity of spatial patterns. Recently, the use of deep learning (DL) for image analysis has increased compared to commonly used machine learning (ML) methods. This paper compares the performance of four algorithms, Random Forest (RF), Support Vector Machine (SVM), Multilayer Perceptron (MLP) and Convolutional Network (CNN), using multi-source data, applying an exhaustive optimisation process of the hyperparameters. The usual approach in the optimisation process of a LULC classification model is to keep the best model in terms of accuracy without analysing the rest of the results. In this study, we have analysed such results, discovering noteworthy patterns in a space defined by the mean and standard deviation of the validation accuracy estimated in a 10-fold cross validation (CV). The point distributions in such a space do not appear to be completely random, but show clusters of points that facilitate the discovery of hyperparameter values that tend to increase the mean accuracy and decrease its standard deviation. RF is not the most accurate model, but it is the less sensitive to changes in hyperparameters. Neural Networks, tend to increase commission and omission errors of the less represented classes because their optimisation lead the model to learn better the most frequent classes. On the other hand, RF and MLP prediction layers are the most accurate from a general qualitative point of view. [ABSTRACT FROM AUTHOR]

This study investigates monthly streamflow modeling at Kale and Durucasu stations in the Black Sea Region of Turkey using remote sensing data. The analysis incorporates key meteorological variables, including air temperature, relative humidity, soil wetness, wind speed, and precipitation. The study also investigates the accuracy of multivariate adaptive regression (MARS) with Kmeans clustering (MARS-Kmeans) by comparing it with single MARS, M5 model tree (M5Tree), random forest regression (RF), multilayer perceptron neural network (MLP). In the first modeling stage, principal component regression is applied to diverse input combinations, both with and without lagged streamflow (Q), resulting in twenty-three and twenty input combinations, respectively. Results demonstrate the critical role of including lagged Q for improved model accuracy, as models without lagged Q exhibit significant performance degradation. The second stage involves a comparative analysis of the MARS-Kmeans model with other machine-learning models, utilizing the best-input combination. MARS-Kmeans, incorporating three clusters, consistently outperforms other models, showcasing superior accuracy in predicting monthly streamflow. [ABSTRACT FROM AUTHOR]

In recent years, machine learning (ML) has become increasingly popular in various fields of activity. Cloud platforms have also grown in popularity, as they offer services that are more secure and accessible worldwide. In this context, cloud-based technologies emerged to support ML, giving rise to the machine learning as a service (MLaaS) concept. However, the clients accessing ML services in order to obtain classification results on private data may be reluctant to upload sensitive information to cloud. The model owners may also prefer not to outsource their models in order to prevent model inversion attacks and to protect intellectual property. The privacy-preserving evaluation of ML models is possible through multi-key homomorphic encryption (MKHE), that allows both the client data and the model to be encrypted under different keys. In this paper, we propose an MKHE evaluation method for decision trees and we extend the proposed method for random forests. Each decision tree is evaluated as a single lookup table, and voting is performed at the level of groups of decision trees in the random forest. We provide both theoretical and experimental evaluations for the proposed method. The aim is to minimize the performance degradation introduced by the encrypted model compared to a plaintext model while also obtaining practical classification times. In our experiments with the proposed MKHE random forest evaluation method, we obtained minimal (less than 0.6%) impact on the main ML performance metrics considered for each scenario, while also achieving reasonable classification times (of the order of seconds). [ABSTRACT FROM AUTHOR]

Background: Despite the clear clinical diagnostic criteria for necrozoospermia in andrology, the fundamental mechanisms underlying it remain elusive. This study aims to profile the lipid composition in seminal plasma systematically and to ascertain the potential of lipid biomarkers in the accurate diagnosis of necrozoospermia. It also evaluates the efficacy of a lipidomics-based random forest algorithm model in identifying necrozoospermia. Methods: Seminal plasma samples were collected from patients diagnosed with necrozoospermia (n = 28) and normozoospermia (n = 28). Liquid chromatography–mass spectrometry (LC–MS) was used to perform lipidomic analysis and identify the underlying biomarkers. A lipid functional enrichment analysis was conducted using the LION lipid ontology database. The top 100 differentially significant lipids were subjected to lipid biomarker examination through random forest machine learning model. Results: Lipidomic analysis identified 46 lipid classes comprising 1267 lipid metabolites in seminal plasma. The top five enriched lipid functions as follows: fatty acid (FA) with ≤ 18 carbons, FA with 16–18 carbons, monounsaturated FA, FA with 18 carbons, and FA with 16 carbons. The top 100 differentially significant lipids were subjected to machine learning analysis and identified 20 feature lipids. The random forest model identified lipids with an area under the curve > 0.8, including LPE(20:4) and TG(4:0_14:1_16:0). Conclusions: LPE(20:4) and TG(4:0_14:1_16:0), were identified as differential lipids for necrozoospermia. Seminal plasma lipidomic analysis could provide valuable biochemical information for the diagnosis of necrozoospermia, and its combination with conventional sperm analysis may improve the accuracy and reliability of the diagnosis. [ABSTRACT FROM AUTHOR]

This study presents a novel procedure and mathematical model employing four artificial intelligence AI algorithms to predict the cumulative degree of soil compaction CDSC during dynamic compaction DC. The four AI algorithms used in this study involve artificial neural network ANN, support vector regression SVR, gradient boosting machine GBM, and random forest RF. Input variables involve the average SPT N value Nini before dynamic compaction, cumulative applied energy normalized with a cross-sectional area of tamper Ea, and the number of the tamper drops Ndrops. Apart from cross-validation with a testing set, additional in situ test data gathered from a different section within the study site are used to assess the generalization capacity of the AI models. In addition, out-of-distribution analyses for the four AI algorithms are conducted in the context of parametric studies. The CDSC prediction performance for the four AI models leads to high prediction metrics of accuracy with the r2 greater than 0.9 for the testing scenario while the r2 of the other AI models is greater than 0.9 when out-of-sample data are considered except for the GBM. The ANN appears to be the best model as the parametric study takes into account out-of-distribution data and suggests a robust relationship between input variables and CDSC that is more coherent with engineering principles for DC. Finally, the ANN model is utilized to develop a new mathematical model for CDSC prediction. [ABSTRACT FROM AUTHOR]