Https://msi.elpub.ru/jour/article/view/1711

Random forests (RF) have long been a widely popular method in medical image analysis. Meanwhile, the closely related gradient boosted trees (GBT) have not become a mainstream tool in medical imaging despite their attractive performance, perhaps due to their computational cost. In this paper, we leverage the recent availability of an efficient open-source GBT implementation to illustrate the GBT method in a corrective learning framework, in application to the segmentation of the caudate nucleus, putamen and hippocampus. The size and shape of these structures are used to derive important biomarkers in many neurological and psychiatric conditions. However, the large variability in deep gray matter appearance makes their automated segmentation from MRI scans a challenging task. We propose using GBT to improve existing segmentation methods. We begin with an existing 'host' segmentation method to create an estimate surface. Based on this estimate, a surface-based sampling scheme is used to construct a set of candidate locations. GBT models are trained on features derived from the candidate locations, including spatial coordinates, image intensity, texture, and gradient magnitude. The classification probabilities from the GBT models are used to calculate a final surface estimate. The method is evaluated on a public dataset, with a 2-fold cross-validation. We use a multi-atlas approach and FreeSurfer as host segmentation methods. The mean reduction in surface distance error metric for FreeSurfer was 0.2 - 0.3 mm, whereas for multi-atlas segmentation, it was 0.1mm for each of caudate, putamen and hippocampus. Importantly, our approach outperformed an RF model trained on the same features (p < 0.05 on all measures). Our method is readily generalizable and can be applied to a wide range of medical image segmentation problems and allows any segmentation method to be used as input.

This study investigates the use of machine learning (ML) models combined with a Synthetic Minority Over-sampling Technique (SMOTE) and its variants to predict perioperative pressure injuries (PIs) in an imbalanced dataset. PIs are a significant healthcare problem, often leading to prolonged hospitalization and increased medical costs. Conventional risk assessment scales are limited in their ability to predict PIs accurately, prompting the exploration of ML techniques to address this challenge.We utilized data from 7,292 patients admitted to a tertiary care hospital in Shanghai between May 2017 and July 2023, with a final dataset of 2,972 patients, including 158 with PIs. Seven ML algorithms-Support Vector Machine (SVM), Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Extra Trees (ET), K-Nearest Neighbors (KNN), and Decision Trees (DT)-were used in conjunction with SMOTE, SMOTE+ENN, Borderline-SMOTE, ADASYN, and GAN to balance the dataset and improve model performance.Results revealed significant improvements in model performance when SMOTE and its variants were used. For instance, the XGBoost model hadan AUC of 0.996 with SMOTE, compared to 0.800 on raw data. SMOTE+ENN and Borderline-SMOTE further enhanced the models' ability to identify minority classes. External validation indicatedthat XGBoost, RF, and ET exhibited the highest stability and accuracy, with XGBoost having an AUC of 0.977. SHAP analysis revealed that factors such as anesthesia grade, age, and serum albumin levels significantly influenced model predictions.In conclusion, integrating SMOTE with ML algorithms effectively addressed a data imbalance and improved the prediction of perioperative PIs. Future work should focus on refining SMOTE techniques and exploring their application to larger, multi-center datasets to enhance the generalizability of these findings, and especially for diseaseswith a lowincidence.

The evaluation of load-settlement behavior of piles is very crucial in meeting the serviceability criteria for pile analysis and design. The most reliable approach for estimating this behavior can be achieved by conducting pile load tests. However, due to the considerable expense and time requirement of such in-situ testing, the load-transfer methods have been used routinely in practice. In this paper, an alternative tree-based machine learning (ML) modeling is explored to predict the load-settlement behavior of axially loaded single piles from cone penetration test (CPT) data. Two variants of tree-based ML models, the random forest (RF) and gradient boosted tree (GBT), are developed in this study to estimate the load-settlement behavior of piles from CPT data (corrected cone tip resistance, qt, and sleeve friction, fs). A database of load-settlement curves of 64 static pile load tests and the corresponding CPT test data were compiled and used for the development of these ML models. The developed RF and GBT models are evaluated based on several statistical criteria. The load-settlement curves for six PLTs predicted using the developed RF and GBT models were compared with the measured data and the load-settlement curves predicted using the conventional load-transfer methods. The results demonstrated the great potential of tree-based ML (RF, GBT) models for predicting the load-settlement behavior of axially loaded piles from CPT data. The comparison clearly shows that the ML models outperformed the conventional load-transfer methods. Amongst the two ML models, the results show that the GBT model outperformed the RF model.

Prediction models to identify individuals at risk of metabolic syndrome who are unlikely to participate in a health intervention program

We aimed to establish a rapid and precise method for identifying and quantifying multi-species vegetable oil (corn oil, olive oil (OLO), soybean oil, and sunflower oil (SUO)) adulterations in camellia oil (CAO), using soft ionization by chemical reaction in transfer–high-resolution mass spectrometry (SICRIT-HRMS) and machine learning methods. The results showed that SICRIT-HRMS could effectively characterize the volatile profiles of pure and adulterated CAO samples, including binary, ternary, quaternary, and quinary adulteration systems. The low m/z region (especially 100–300) exhibited importance to oil classification in multiple feature-selection methods. For qualitative detection, binary classification models based on convolutional neural networks (CNN), Random Forest (RF), and gradient boosting trees (GBT) algorithms showed high accuracies (98.70–100.00%) for identifying CAO adulteration under no dimensionality reduction (NON), principal component analysis (PCA), and uniform manifold approximation and projection (UMAP) strategies. The RF algorithm exhibited relatively high accuracy (96.25–99.45%) in multiclass classification. Moreover, the five models, including CNN, RF, support vector machines (SVM), logistic regression (LR), and GBT, exhibited different performances in distinguishing pure and adulterated CAO. Among 1093 blind oil samples, under NON, PCA, and UMAP: 10, 5, and 67 samples were misclassified by CNN model; 6, 7, and 41 samples were misclassified by RF model; 8, 9, and 82 samples were misclassified by SVM model; 17, 18, and 78 samples were misclassified by LR model; 7, 9, and 43 samples were misclassified by GBT model. For quantitative prediction, the PCA-CNN model performed optimally in predicting adulteration levels in CAO, especially with respect to OLO and SUO, exhibiting a high coefficient of determination for calibration (, 0.9664–0.9974) and coefficient of determination for prediction (, 0.9599–0.9963) values, low root mean square error of calibration (RMSEC, 0.9–5.3%) and root mean square error of prediction (RMSEP, 1.1–5.8%) values, and RPD (5.0–16.3) values greater than 3.0. These results indicate that SICRIT-HRMS combined with machine learning can rapidly and accurately identify and quantify multi-species vegetable oil adulterations in CAO, which provides a reference for developing non-targeted and high-throughput detection methods in edible oil authenticity.

This study conducts an extensive comparative analysis of computational intelligence approaches aimed at predicting the compressive strength (CS) of concrete, utilizing two non-destructive testing (NDT) methods: the rebound hammer (RH) and the ultrasonic pulse velocity (UPV) test. In the ensemble learning approach, the six most popular algorithms (Adaboost, CatBoost, gradient boosting tree (GBT), random forest (RF), stacking, and extreme gradient boosting (XGB)) have been used to develop the prediction models of CS of concrete based on NDT. The ML models have been developed using a total of 721 samples, of which 111 were cast in the laboratory, 134 were obtained from in-situ testing, and the other samples were gathered from the literature. Among the three categories of analytical models—RH models, UPV models, and combined RH and UPV models; seven, ten, and thirteen models have been used respectively. AdaBoost, CatBoost, GBT, RF, Stacking, and XGB models have been used to improve the accuracy and dependability of the analytical models. The RH-M5, UPV-M6, and C-M6 (combined UPV and RH model) models were found with highest performance level amongst all the analytical models. The MAPE value of XGB was observed to be 84.37%, 83.24%, 77.33%, 59.46%, and 81.08% lower than AdaBoost, CatBoost, GBT, RF, and stacking, respectively. The performance of XGB model has been found best than other soft computing techniques and existing traditional predictive models.

Accurate and reliable knowledge about grassland distribution is essential for farmers, stakeholders, and government to effectively manage grassland resources from agro-economical and ecological perspectives. This study developed a novel pixel-based grassland classification approach using three supervised machine learning (ML) algorithms, which were assessed in the province of Manitoba, Canada. The grassland classification process involved three stages: (1) to distinguish between vegetation and non-vegetation covers, (2) to differentiate grassland from non-grassland landscapes, and (3) to identify three specific grassland classes (tame, native, and mixed grasses). Initially, this study investigated different satellite data, such as Sentinel-1 (S1), Sentinel-2 (S2), and Landsat 8 and 9, individually and combined, using the random forest (RF) method, with the best performance at the first two steps achieved using a combination of S1 and S2. The combination was then utilized to conduct the first two steps of classification using support vector machine (SVM) and gradient tree boosting (GTB). In step 3, after filtering out non-grassland pixels, the performance of RF, SVM, and GTB classifiers was evaluated with combined S1 and S2 data to distinguish different grassland types. Eighty-nine multitemporal raster-based variables, including spectral bands, SAR backscatters, and digital elevation models (DEM), were input for ML models. RF had the highest classification accuracy at 69.96% overall accuracy (OA) and a Kappa value of 0.55. After feature selection, the variables were reduced to 61, increasing OA to 72.62% with a Kappa value of 0.58. GTB ranked second, with its OA and Kappa values improving from 67.69% and 0.50 to 72.18% and 0.58 after feature selection. The impact of raster data quality on grassland classification accuracy was assessed through multisensor image fusion. Grassland classification using the Hue, Saturation, and Value (HSV) fused images showed higher OA (59.18%) and Kappa values (0.36) than the Brovey Transform (BT) and non-fused images. Finally, a web map was created to show grassland results within the Soil Landscapes of Canada (SLC) polygons, relating soil landscapes to grassland distribution and providing valuable information for decision-makers and researchers. Future work may include extending the current methodology by considering other influential variables, like meteorological parameters or soil properties, to create a comprehensive grassland inventory across the whole Prairie ecozone of Canada.

Utilisation of machine learning algorithms to predict relapse risk following allogeneic stem cell transplantation (alloSCT) for patients with myeloid malignancies

Educational data mining (EDM) is the techniques used to discover the knowledge from student's data .it is used to improve the students' performance and teachers' performances as well. Since Machine learning (ML) models deals with numeric data, preprocessing of the categorical data is a must step to transform such data into accepted types by ML models. Categorical data is further divided into nominal and ordinal attributes in the dataset. The used data set was collected by using a learner activity tracker tool, which called experience API (xAPI). The purposed was to monitor the behaviors of students to evaluate the features that may impact the student performance (Amrieh, Hamtini &amp; Aljarah, 2015). The dataset includes 480 student records and 16 features. The features are eitrher numeric or catgorical. In this paper, we study the effect of encoding some non-ordinal features as one-hot (dummy variables) on the students' performance prediction accuracy. We used techniques form ensemble methods such as Random Forest Trees, Boosting methods specifically namely gradient Boosted trees (GBT), and support vector machines. Also, we compared the performance of Random forest and Gradient boosted trees. We achieve a better result of 81% using random forest classifier. GBT has approximately same performance in all cases. SVM accuracy improved when used dummy variables. In addition, we show the importance of the behavioral features as they increase the model's accuracy. Random Forest outperformed other models considered in this study. At the end, the encoding methods as preprocessing steps can affect the performance accuracy of the models.

Considering the form diversity of tree species composition in the Bagong Mountain National Forest Park of China, we mapped tree species utilizing Machine Learning Algorithms (support vector machines (SVM) and random forest (RF) classifiers) based on the OHS-2 hyperspectral satellite image by different datasets which combined spectral information and hyperspectral-derived vegetation indices (VIs) for improving tree species classification and explored the best performance of them. To verify the improvement, the results of physically-based spectral classifiers (spectral angle mapper (SAM) and maximum likelihood (ML) classifiers) were applied to compare with the results of machine learning algorithms. The results indicated an overall accuracy of 94.01%, 96.08%, 82.9% and 79.3% for SVM, RF, SAM and ML classifiers of the best performance using different datasets. Highest accuracies resulted from two machine learning algorithms classifiers; SVM and RF compared to SAM and ML classifiers. Although SVM outperformed RF when using all hyperspectral bands and VIs, the overall accuracy of the RF classifier is higher when compared to the SVM classifier using VIs combined selected features. Meanwhile, the RF classifier performed better than SVM after removing the redundancy of spectral data in training samples. Moreover, the machine learning algorithms successfully classified a small number of tree species (Cedrus deodara and Pterocarya stenoptera C. DC.) in the study area, but the physical spectroscopy-based method failed to classify these species. Such integration strategy improved the effectiveness of enhancing the accuracy of tree species classification and mapping their distribution on broad spatial and temporal scales using machine learning algorithms and hyperspectral imagery.

PURPOSEThe optimal characteristics among patients with breast cancer to recommend neoadjuvant chemotherapy is an active area of clinical research. We developed and compared several approaches to developing prediction models for pathologic complete response (pCR) among patients with breast cancer in Alberta.METHODSThe study included all patients with breast cancer who received neoadjuvant chemotherapy in Alberta between 2012 and 2014 identified from the Alberta Cancer Registry. Patient, tumor, and treatment data were obtained through primary chart review. pCR was defined as no residual invasive tumor at surgical excision in breast or axilla. Two types of prediction models for pCR were built: (1) expert model: variables selected on the basis of oncologists' opinions and (2) data-driven model: variables selected by trained machine. These model types were fit using logistic regression (LR), random forests (RF), and gradient-boosted trees (GBT). We compared the models using area under the receiver operating characteristic curve and integrated calibration index, and internally validated using bootstrap resampling.RESULTSA total of 363 cases were included in the analyses, of which 86 experienced pCR. The RF and GBT fits yielded higher optimism-corrected area under the receiver operating characteristic curves compared with LR for the expert (RF: 0.70; GBT: 0.69; LR: 0.65) and data-driven models (RF: 0.71; GBT: 0.68; LR: 0.64). The LR fit yielded the lowest integrated calibration indices for the expert (LR: 0.037; GBT: 0.05; RF: 0.10) and data-driven models (LR: 0.026; GBT: 0.06; RF: 0.099).CONCLUSIONOur models demonstrated predictive ability for pCR using routinely collected clinical and demographic variables. We show that machine learning fit methods can be used to optimize models for pCR prediction. We also show that additional variables beyond clinical expertise do not considerably improve predictive ability and may not be of value on the basis of the burden of data collection.

Purpose This study aims to develop and validate a machine learning-based risk prediction model for 1-year cardiac mortality in acute myocardial infarction (AMI) patients, utilizing a multidimensional dataset that integrates traditional risk factors, environmental determinants, and angiographic findings. Methods Data was sourced from a multicenter prospective cohort study (NCT 03297164) along with electronic medical records from a large tertiary hospital, covering the period from January 1, 2017, to July 28, 2022. The patient population from the tertiary hospital constituted the derivation cohort, while patients from other participating centers served as the external validation cohort. Six machine learning algorithms — logistic regression, XGBoost, LightGBM (LGBM), Gradient Boosting, random forest, and decision tree — were applied, with feature selection performed using the SelectFromModel method. The best-performing model underwent both internal and external validation, with its predictive performance compared to the GRACE score. Result The derivation cohort was divided into training and testing cohorts. In the training cohort, six machine learning algorithms were employed to screen variables and construct models using five-fold cross-validation (Figure 2A). All models achieved an area under the receiver operating characteristic curve (AUC) &amp;gt; 0.8 (Figure 2B). Delong's test revealed no significant differences in AUC among the algorithms. Considering other evaluation metrics and brier score, the LGBM (AUC 0.85) model was selected, incorporating the following variables: age, body mass index, smoking history, temperature, PM2.5, Gensini score, weighted diffuse lesion value, left ventricular end-diastolic diameter, ejection fraction, serum creatinine, cardiac troponin I, glucose, low-density lipoprotein cholesterol, B-type natriuretic peptide, and D-dimmer. External validation of the LGBM model yielded an AUC of 0.80, while the GRACE score achieved AUCs of 0.73 and 0.74 in internal testing and external validation, respectively (Figure 2C). Predicted LGBM scores in the training cohort were stratified into ten groups by decile. Patients were categorized into low-, intermediate-, and high-risk groups based on the actual incidence of cardiac death (Figure 2D). Using these cut-off values, risk stratification was applied to both the internal testing cohort and the external validation cohort, and the actual event rates were calculated (Figure E). Logistic and Cox regression analyses demonstrated significant differences between risk groups (P &amp;lt; 0.001, Figures F-I), with the LGBM-based stratification outperforming that based on the GRACE score. Conclusion Based on recent large-scale multicenter cohort data and incorporating environmental factors along with angiographic findings, this study developed and validated a novel prediction model for 1-year cardiac death in AMI patients, which demonstrated superior performance compared to the GRACE score.Flow chart Main Findings

Water quality assessment plays a crucial role in various aspects, including human health, environmental impact, agricultural productivity, and industrial processes. Machine learning (ML) algorithms offer the ability to automate water quality evaluation and allow for effective and rapid assessment of parameters associated with water quality. This article proposes an ML-based classification model for water quality prediction. The model was tested with 14 ML algorithms and considers 20 features that represent various substances present in water samples and their concentrations. The dataset used in the study comprises 7,996 samples, and the model development involves several stages, including data preprocessing, Yeo–Johnson transformation for data normalization, principal component analysis (PCA) for feature selection, and the application of the synthetic minority over-sampling technique (SMOTE) to address class imbalance. Performance metrics, such as accuracy, precision, recall, and F1 score, are provided for each algorithm with and without SMOTE. LightGBM, XGBoost, CatBoost, and Random Forest were identified as the best-performing algorithms. XGBoost achieved the highest accuracy of 96.31% without SMOTE and had a precision of 0.933. The application of SMOTE enhanced the performance of CatBoost. These findings provide valuable insights for ML-based water quality assessment, aiding researchers and professionals in decision-making and management.

Advanced hydrogeochemical facies classification: A comparative analysis of Machine Learning models with SMOTE in the Tawi basin

Credit card fraud remains a significant challenge, with financial losses and consumer protection at stake. This study addresses the need for practical, real-time fraud detection methodologies. Using a Kaggle credit card dataset, I tackle class imbalance using the Synthetic Minority Oversampling Technique (SMOTE) to enhance modeling efficiency. I compare several machine learning algorithms, including Logistic Regression, Linear Discriminant Analysis, K-nearest Neighbors, Classification and Regression Tree, Naive Bayes, Support Vector, Random Forest, XGBoost, and Light Gradient-Boosting Machine to classify transactions as fraud or genuine. Rigorous evaluation metrics, such as AUC, PRAUC, F1, KS, Recall, and Precision, identify the Random Forest as the best performer in detecting fraudulent activities. The Random Forest model successfully identifies approximately 92% of transactions scoring 90 and above as fraudulent, equating to a detection rate of over 70% for all fraudulent transactions in the test dataset. Moreover, the model captures more than half of the fraud in each bin of the test dataset. SHAP values provide model explainability, with the SHAP summary plot highlighting the global importance of individual features, such as “V12” and “V14”. SHAP force plots offer local interpretability, revealing the impact of specific features on individual predictions. This study demonstrates the potential of machine learning, particularly the Random Forest model, for real-time credit card fraud detection, offering a promising approach to mitigate financial losses and protect consumers.