Extra Trees Model Research Articles

A machine learning algorithm to explore the drivers of carbon emissions in Chinese cities

As the world’s largest energy consumer and carbon emitter, the task of carbon emission reduction is imminent. In order to realize the dual-carbon goal at an early date, it is necessary to study the key factors affecting China’s carbon emissions and their non-linear relationships. This paper compares the performance of six machine learning algorithms to that of traditional econometric models in predicting carbon emissions in China from 2011 to 2020 using panel data from 254 cities in China. Specifically, it analyzes the comparative importance of domestic economic, external economic, and policy uncertainty factors as well as the nonparametric relationship between these factors and carbon emissions based on the Extra-trees model. Results show that energy consumption (ENC) remains the root cause of increased carbon emissions among domestic economic factors, although government intervention (GOV) and digital finance (DIG) can significantly reduce it. Next, among the external economic and policy uncertainty factors, foreign direct investment (FDI) and economic policy uncertainty (EPU) are important factors influencing carbon emissions, and the partial dependence plots (PDPs) confirm the pollution haven hypothesis and also reveal the role of EPU in reducing carbon emissions. The heterogeneity of factors affecting carbon emissions is also analyzed under different city sizes, and it is found that ENC is a common driving factor in cities of different sizes, but there are some differences. Finally, appropriate policy recommendations are proposed by us to help China move rapidly towards a green and sustainable development path.

Video source identification using machine learning: A case study of 16 instant messaging applications

In recent years, there has been a notable increase in the prevalence of cybercrimes related to video content, including the distribution of illegal videos and the sharing of copyrighted material. This has led to the growing importance of identifying the source of video files to trace the owner of the files involved in the incident or identify the distributor. Previous research has concentrated on revealing the device (brand and/or model) that “originally” created a video file. This has been achieved by analysing the pattern noise generated by the image sensor in the camera, the storage structural features of the file, and the metadata patterns. However, due to the widespread use of mobile environments, instant messaging applications (IMAs) such as Telegram and Wire have been utilized to share illegal videos, which can result in the loss of information from the original file due to re-encoding at the application level, depending on the transmission settings. Consequently, it is necessary to extend the scope of existing research to identify the various applications that are capable of re-encoding video files in transit. Furthermore, it is essential to determine whether there are features that can be leveraged to distinguish them from the source identification perspective. In this paper, we propose a machine learning-based methodology for classifying the source application by extracting various features stored in the storage format and internal metadata of video files. To conduct this study, we analyzed 16 IMAs that are widely used in mobile environments and generated a total of 1974 sample videos, taking into account both the transmission options and encoding settings offered by each IMA. The training and testing results on this dataset indicate that the ExtraTrees model achieved an identification accuracy of approximately 99.96 %. Furthermore, we developed a proof-of-concept tool based on the proposed method, which extracts the suggested features from videos and queries a pre-trained model. This tool is released as open-source software for the community.

A machine learning approach towards assessing consistency and reproducibility: an application to graft survival across three kidney transplantation eras.

In South Africa, between 1966 and 2014, there were three kidney transplant eras defined by evolving access to certain immunosuppressive therapies defined as Pre-CYA (before availability of cyclosporine), CYA (when cyclosporine became available), and New-Gen (availability of tacrolimus and mycophenolic acid). As such, factors influencing kidney graft failure may vary across these eras. Therefore, evaluating the consistency and reproducibility of models developed to study these variations using machine learning (ML) algorithms could enhance our understanding of post-transplant graft survival dynamics across these three eras. This study explored the effectiveness of nine ML algorithms in predicting 10-year graft survival across the three eras. We developed and internally validated these algorithms using data spanning the specified eras. The predictive performance of these algorithms was assessed using the area under the curve (AUC) of the receiver operating characteristics curve (ROC), supported by other evaluation metrics. We employed local interpretable model-agnostic explanations to provide detailed interpretations of individual model predictions and used permutation importance to assess global feature importance across each era. Overall, the proportion of graft failure decreased from 41.5% in the Pre-CYA era to 15.1% in the New-Gen era. Our best-performing model across the three eras demonstrated high predictive accuracy. Notably, the ensemble models, particularly the Extra Trees model, emerged as standout performers, consistently achieving high AUC scores of 0.95, 0.95, and 0.97 across the eras. This indicates that the models achieved high consistency and reproducibility in predicting graft survival outcomes. Among the features evaluated, recipient age and donor age were the only features consistently influencing graft failure throughout these eras, while features such as glomerular filtration rate and recipient ethnicity showed high importance in specific eras, resulting in relatively poor historical transportability of the best model. Our study emphasises the significance of analysing post-kidney transplant outcomes and identifying era-specific factors mitigating graft failure. The proposed framework can serve as a foundation for future research and assist physicians in identifying patients at risk of graft failure.

Toward a globally lunar calendar: a machine learning-driven approach for crescent moon visibility prediction

This paper presents a comprehensive approach to harmonizing lunar calendars across different global regions, addressing the long-standing challenge of variations in new crescent Moon sightings that mark the beginning of lunar months. We propose a machine learning (ML)-based framework to predict the visibility of the new crescent Moon, representing a significant advancement toward a globally unified lunar calendar. Our study utilized a dataset covering various countries globally, making it the first to analyze all 12 lunar months over a span of 13 years. We applied a wide array of ML algorithms and techniques. These techniques included feature selection, hyperparameter tuning, ensemble learning, and region-based clustering, all aimed at maximizing the model’s performance. The overall results reveal that the gradient boosting (GB) model surpasses all other models, achieving the highest F1 score of 0.882469 and an area under the curve (AUC) of 0.901009. However, with selected features identified through the ANOVA F-test and optimized parameters, the Extra Trees model exhibited the best performance with an F1 score of 0.887872, and an AUC of 0.906242. We expanded our analysis to explore ensemble models, aiming to understand how a combination of models might boost predictive accuracy. The Ensemble Model exhibited a slight improvement, with an F1 score of 0.888058 and an AUC of 0.907482. Additionally, the geographical segmentation of the dataset enhanced predictive performance in certain areas, such as Africa and Asia. In conclusion, ML techniques can provide efficient and reliable tool for predicting the new crescent Moon visibility that would support the decisions of marking the beginning of new lunar months.

VoiceAD: A multilingual speech based classifier for early detection of Alzheimers disease

Alzheimer's Disease (AD) causes memory decline beyond normal aging and is the most common cause of dementia. Early detection and intervention can delay the progression of AD and alleviate symptoms. However, high expense, insensitivity, and inconvenience are among various deterrents to traditional diagnosis methods. Dementia impacts more than 55 million people globally, with this number set to double every 20 years. Much of this increase will be in low- and middle-income countries which often lack the resources for early diagnosis. Recently speech has emerged as an effective digital biomarker for AD detection. However, most studies in this area are limited to English speech due to lack of suitable data in other languages. This study aims to construct a speech-based classifier (VoiceAD) that can be trained on English data and still maintain its performance on other languages. Acoustic features were extracted from English, Spanish, and Greek data from DementiaBank with the eGeMAPS v2.0 feature set. Subsequently, an Extra Trees model was employed for classification. VoiceADs performance for English and Greek surpassed that of the baseline model and model incorporating only demographic data while the Spanish performance was lower due to task effect. In addition, VoiceADs satisfactory generalizability and interpretability positions it as a promising candidate for clinical applications. It can be implemented as a free screening tool such as a mobile application to monitor AD risk. VoiceADs crosslingual applicability can its enable worldwide deployment, especially in settings where traditional methods of diagnosis are less attainable.

Multimodal integration of radiology and pathology signatures for distinguishing between aldosterone-producing adenomas and nonfunctional adrenal adenomas.

To develop and validate a nomogram combining radiomics and pathology features to distinguish between aldosterone-producing adenomas (APAs) and nonfunctional adrenal adenomas (NF-AAs). Consecutive patients diagnosed with adrenal adenomas via computed tomography (CT) or pathologic analysis between January 2011 and November 2022 were eligible for inclusion in this retrospective study. CT images and hematoxylin & eosin-stained slides were used for annotation and feature extraction. The selected radiomics and pathology features were used to develop a risk model using various machine learning models, and the area under the receiver operating characteristic curve (AUC) was determined to evaluate diagnostic performance. The predicted results from radiomics and pathology features were combined and visualized using a nomogram. A total of 211 patients (APAs, n = 59; NF-AAs, n = 152) were included in this study, with patients randomly divided into either the training set or the testing set at a ratio of 8:2. The ExtraTrees model yielded a sensitivity of 0.818, a specificity of 0.733, and an accuracy of 0.756 (AUC = 0.817; 95% confidence interval [CI]: 0.675-0.958) in the radiomics testing set and a sensitivity of 0.999, a specificity of 0.842, and an accuracy of 0.867 (AUC = 0.905, 95% CI: 0.792-1.000) in the pathology testing set. A nomogram combining radiomics and pathology features demonstrated a strong performance (AUC = 0.912; 95% CI: 0.807-1.000). A nomogram combining radiomics and pathology features demonstrated strong predictive accuracy and discrimination capability. This model may help clinicians to distinguish between APAs and NF-AAs.

Automated Prediction of Proximal Middle Cerebral Artery Occlusions in Noncontrast Brain Computed Tomography.

Early identification of large vessel occlusion (LVO) in patients with ischemic stroke is crucial for timely interventions. We propose a machine learning-based algorithm (JLK-CTL) that uses handcrafted features from noncontrast computed tomography to predict LVO. We included patients with ischemic stroke who underwent concurrent noncontrast computed tomography and computed tomography angiography in seven hospitals. Patients from 5 of these hospitals, admitted between May 2011 and March 2015, were randomly divided into training and internal validation (9:1 ratio). Those from the remaining 2 hospitals, admitted between March 2021 and September 2021, were designated for external validation. From each noncontrast computed tomography scan, we extracted differences in volume, tissue density, and Hounsfield unit distribution between bihemispheric regions (striatocapsular, insula, M1-M3, and M4-M6, modified from the Alberta Stroke Program Early Computed Tomography Score). A deep learning algorithm was used to incorporate clot signs as an additional feature. Machine learning models, including ExtraTrees, random forest, extreme gradient boosting, support vector machine, and multilayer perceptron, as well as a deep learning model, were trained and evaluated. Additionally, we assessed the models' performance after incorporating the National Institutes of Health Stroke Scale scores as an additional feature. Among 2919 patients, 83 were excluded. Across the training (n=2463), internal validation (n=275), and external validation (n=95) datasets, the mean ages were 68.5±12.4, 67.6±13.8, and 67.9±13.6 years, respectively. The proportions of men were 57%, 53%, and 59%, with LVO prevalences of 17.0%, 16.4%, and 26.3%, respectively. In the external validation, the ExtraTrees model achieved a robust area under the curve of 0.888 (95% CI, 0.850-0.925), with a sensitivity of 80.1% (95% CI, 72.0-88.1) and a specificity of 88.6% (95% CI, 84.7-92.5). Adding the National Institutes of Health Stroke Scale score to the ExtraTrees model increased sensitivity (from 80.1% to 92.1%) while maintaining specificity. Our algorithm provides reliable predictions of LVO using noncontrast computed tomography. By enabling early LVO identification, our algorithm has the potential to expedite the stroke workflow.

Tree-based Machine Learning Ensembles and Feature Importance Approach for the Identification of Intrusions in UNR-IDD Dataset

Detection of intrusions from network data with the use of machine learning techniques has gained great attention in the past decades. One of the key problems in the network security domain is the availability of representative datasets for testing and evaluation purposes. Despite several efforts by researchers to release datasets that can be used for benchmarking attack detection models, some of the released datasets still suffer from one limitation or the other. Thus, some researchers at the University of Nevada released a dataset named UNR-IDD dataset which was argued to be free from some of the limitations of the past datasets. This study proposed Tree-based ensemble approaches for building binary intrusion identification models from the UNR-IDD dataset. Decision Tree algorithms are used as base classifiers in the Extra Trees, Random Forest and AdaBoost-based intrusion detection models. The results of the experimental analyses carried out indicated that the three ensembles performed excellently when feature selection was used compared to when all features were applied. For instance, Extra Trees model achieved an accuracy of 0.97, precision of 0.98, recall of 0.98 and f1-score of 0.98. Similarly, Random Forest model achieved an accuracy of 0.98, precision of 0.98, recall of 0.99 and f1-score of 0.98. Adaboost-based model had an accuracy of 0.96, precision of 0.96, recall of 0.99 and f1-score of 0.98. It was deduced that Random Forest intrusion classification model achieved slight overall best results when compared to the other models built. It is concluded that the three homogeneous ensemble models achieved very promising results while feature importance was used as attribute selection method.

Effective approach to assess higher heating value of biomass from ultimate and proximate analysis

AbstractFor proper design and operation of biomass‐based energy systems, it is important to determine the higher heating value (HHV) of biomass. In this paper, two machine learning (ML) approaches, namely extra trees (ET) and least squares support vector machine (LSSVM), are used to predict the value of HHV associated with biofuels. The data required for HHV calculation, including proximate and ultimate analyses datasets, were collected from the literature. The performances of these two ML approaches for predicting biomass HHV were then compared with other smart models available in the literature. Even though the available empirical models can predict the biomass HHV with acceptable precision, it was found that our proposed ML techniques have a superior performance based on the error analysis; the proposed approaches also consider all key biomass characteristics in the developed models. In addition, the ET model proved to be slightly more accurate compared to the LSSVM model. Additionally, the developed proximate‐based ET model showed better performance compared to the ultimate‐based ET model. The most influential parameters in the developed ET models for the proximate and ultimate approaches were determined to be ash fraction and carbon fraction, respectively. Finally, it was concluded that the smart modelling techniques can be utilized as a robust and reliable alternative predictive methodology to replace direct laboratory measurement of the biomass HHV.

Diabetes Prediction: Optimization of Machine Learning through Feature Selection and Dimensionality Reduction

Diabetes, a pervasive global health concern, presents diagnostic challenges due to its nuanced onset and far-reaching implications. Traditional diagnostic approaches, reliant on time-consuming assessments, necessitate a paradigm shift towards more efficient methodologies. In response, this study introduces a diagnostic support system leveraging the power of optimized machine learning algorithms. Addressing class imbalance within a dataset comprising 768 records, our methodology intricately weaves together feature selection, dimensionality reduction techniques, and grid search optimization. Specifically, the Extra Trees model, fine-tuned via grid search, emerges as the most potent, showcasing remarkable performance metrics: an accuracy score of 92.5%, an F1-score of 93.7%, and an AUC-ROC of 92.47%. These findings underscore the pivotal role of machine learning in reshaping diabetes diagnosis, offering transformative possibilities for global healthcare enhancement.

Predict the surface free energy of Ti/Zr/V chemical conversion coating on nickel by machine learning approach: Trees vs. neurons

The surface properties of chemical conversion treated materials have received increasing attention and surface free energy (SFE) is an important indicator of it. In this study, an environment-friendly Ti/Zr/V chemical conversion coating was deposited on the nickel and Artificial Neural Network (ANN), Random Forest (RF), and Extra Trees (ET) models were applied to further investigate the relationship between the SFE and the compositions of the conversion bath. The accuracy of SFE prediction of three machine learning models was quantified, and the obtained results reveal that the correlation coefficients for the ANN, RF, and ET models are 0.958, 0.979, and 0.963 for the prediction of dispersion components and 0.964, 0.942 and 0.921 for the prediction of polar components, respectively, and the ANN model shows more outstanding performance in general. The errors in the prediction of the ANN model for the polar and dispersion components of SFE are 0.336 and 0.421, respectively, which are smaller than the extended uncertainties obtained from conventional experimental methods. The obtained machine learning models can accurately predict the variations of SFE for modest changes in the conversion bath components, thus linking the formation mechanism of coating with the variations of SFE.

An empirical assessment of ML models for 5G network intrusion detection: A data leakage-free approach

This paper thoroughly compares thirteen unique Machine Learning (ML) models utilized for Intrusion detection systems (IDS) in a meticulously controlled environment. Unlike previous studies, we introduce a novel approach that meticulously avoids data leakage, enhancing the reliability of our findings. The study draws upon a comprehensively labeled 5G-NIDD dataset covering a broad spectrum of network behaviors, from benign real-user traffic to various attack scenarios. Our data preprocessing and experimental design have been carefully structured to eradicate any data leakage, a standout feature of our methodology that significantly improves the robustness and dependability of our results compared to prior studies. The ML models are evaluated using various performance metrics, including accuracy, precision, recall, F1-score, ROC AUC, and execution time. Our results reveal that the K-Nearest Neighbors model is superior in accuracy and ROC AUC, while the Voting Classifier stands out in precision and F1-score. Decision Tree, Bagging, and Extra Trees models exhibit strong recall scores. In contrast, the AdaBoost model falls short across all assessed metrics. Despite displaying only modest performance on other metrics, the Naive Bayes model excels in computational efficiency, offering the quickest execution time. This paper emphasizes the importance of understanding various ML models' distinct strengths, drawbacks, and trade-offs for network intrusion detection. It highlights that no single model is universally superior, and the choice hinges on the nature of the dataset, specific application requirements, and the computational resources available.

Polymer design using machine learning: A quest for high glass transition temperature

This research introduces an advanced data-centric framework tailored for polymer design. Developing a machine learning (ML) model utilizing molecular descriptors, the study aims to forecast the glass transition temperature of polymers. Employing the Retrosynthetically Interesting Chemical Substructures (BRICS) technique, a set of 10,000 novel polymers is synthesized, followed by predicting their glass transition temperature (Tg) values using a pre-existing ML model trained for this purpose. Gradient Boosting and Extra Trees models outperform K Neighbors and Random Forest (KNN) models in predicting performance parameters with R2 scores of 0.99 and 0.99, and lower root mean squared error values of 24 and 32 K, respectively. The selection of polymers is done on the basis of glass transition temperature, those with higher glass transition temperature are selected. We evaluate the ease of synthetic accessibilities of the selected polymers alongwith their structural variations which indicates a diverse range among them. Through the successful identification and enhancement of novel polymers, these innovative methods greatly enhance the possibilities of discovering superior materials tailored for cutting-edge applications.

General Model for Predicting Response of Gas-Sensitive Materials to Target Gas Based on Machine Learning.

Gas sensors play a crucial role in various industries and applications. In recent years, there has been an increasing demand for gas sensors in society. However, the current method for screening gas-sensitive materials is time-, energy-, and cost-consuming. Consequently, an imperative exists to enhance the screening efficiency. In this study, we proposed a collaborative screening strategy through integration of density functional theory and machine learning. Taking zinc oxide (ZnO) as an example, the responsiveness of ZnO to the target gas was determined quickly on the basis of the changes in the electronic state and structure before and after gas adsorption. In this work, the adsorption energy and electronic and structural characteristics of ZnO after adsorbing 24 kinds of gases were calculated. These computed features served as the basis for training a machine learning model. Subsequently, various machine learning and evaluation algorithms were utilized to train the fast screening model. The importance of feature values was evaluated by the AdaBoost, Random Forest, and Extra Trees models. Specifically, charge transfer was assigned importance values of 0.160, 0.127, and 0.122, respectively, ranking as the highest among the 11 features. Following closely was the d-band center, which was presumed to exert influence on electrical conductivity and, consequently, adsorption properties. With 5-fold cross-validation using the Extra Tree accuracy, the 24-sample data set achieved an accuracy of 88%. The 72-sample data set achieved an accuracy of 78% using multilayer perceptron after 5-fold cross-validation, with both data sets exhibiting low standard deviations. This verified the accuracy and reliability of the strategy, showcasing its potential for rapidly screening a material's responsiveness to the target gas.

Air pollution and diabetes mellitus: Association and validation in a desert area in China.

Despite the growing evidence pointing to the detrimental effects of air pollution on diabetes mellitus (DM), the relationship remains poorly explored, especially in desert-adjacent areas characterized by high aridity and pollution. We conducted a cross-sectional study with health examination data from over 2.9 million adults in two regions situated in the southern part of the Taklamakan Desert, China. We assessed three-year average concentrations (2018-2020) of particulate matter (PM1, PM2.5, and PM10), carbon monoxide (CO), nitrogen dioxide (NO2), and sulfur dioxide (SO2) through a space-time extra-trees model. After adjusting for various covariates, we employed generalized linear mixed models to evaluate the association between exposure to air pollutants and DM. The odds ratios for DM associated with a 10 µg/m3 increase in PM1, PM2.5, PM10, CO, and NO2 were 1.898 (95% CI: 1.741, 2.070), 1.07 (95% CI: 1.053, 1.086), 1.013 (95% CI: 1.008, 1.018), 1.009 (95% CI: 1.007, 1.011), and 1.337 (95% CI: 1.234, 1.449), respectively. Notably, men, individuals aged ≥50 years, those with lower educational attainment, nonsmokers, and those not engaging in physical exercise displayed more susceptible to the adverse effects of air pollution. Multiple sensitivity analyses confirmed the stability of these findings. Our study provides robust evidence of a correlation between prolonged exposure to air pollution and the prevalence of DM among individuals living in the desert-adjacent areas. This research contributes to the expanding knowledge on the relationship between air pollution exposure and DM prevalence in desert-adjacent areas.

Advanced Machine Learning Techniques for Liver Tumor Classification in MRI Imaging

In this research into liver tumor categorization within MRI images, diverse machine learning methodologies were scrutinized for their efficacy. The study delved into the integration of shape and texture features, aiming to bolster classification accuracy. Among the algorithms explored, the Extra Trees model emerged as the most promising contender, exhibiting superior performance compared to its counterparts. Leveraging the distinctive capabilities of the Extra Trees model, the study underscored its effectiveness in accurately categorizing liver tumors. This highlights its potential to enhance diagnostic precision in clinical contexts. Through rigorous experimentation and analysis, the research elucidated the significance of incorporating shape and texture features into machine learning frameworks for improved tumor classification. The findings not only contribute to advancing the field of medical imaging but also underscore the importance of leveraging innovative methodologies to address healthcare challenges. Overall, the study sheds light on the promising prospects of employing advanced machine learning techniques in medical imaging for more accurate and efficient diagnosis of liver tumors.

Remaining Useful Life Estimation Framework for the Main Bearing of Wind Turbines Operating in Real Time

The prognosis of wind turbine failures in real operating conditions is a significant gap in the academic literature and is essential for achieving viable performance parameters for the operation and maintenance of these machines, especially those located offshore. This paper presents a framework for estimating the remaining useful life (RUL) of the main bearing using regression models fed operational data (temperature, wind speed, and the active power of the network) collected by a supervisory control and data acquisition (SCADA) system. The framework begins with a careful data filtering process, followed by creating a degradation profile based on identifying the behavior of temperature time series. It also uses a cross-validation strategy to mitigate data scarcity and increase model robustness by combining subsets of data from different available turbines. Support vector, gradient boosting, random forest, and extra trees models were created, which, in the tests, showed an average of 20 days in estimating the remaining useful life and presented mean absolute error (MAE) values of 0.047 and mean squared errors (MSE) of 0.012. As its main contributions, this work proposes (i) a robust and effective regression modeling method for estimating RUL based on temperature and (ii) an approach for dealing with a lack of data, a common problem in wind turbine operation. The results demonstrate the potential of using these forecasts to support the decision making of the teams responsible for operating and maintaining wind farms.

Exposure to ambient air pollution and metabolic dysfunction-associated fatty liver disease: Findings from over 2.7 million adults in Northwestern China

Ambient air pollutants exposures may lead to aggravated Metabolic Dysfunction-Associated Fatty Liver Disease (MAFLD). However, there is still a scarcity of empirical studies that have rigorously estimated this association, especially in regions where air pollution is severe. To fill in the literature gap, we conducted a cross-sectional study involving 2711,207 adults living in five regions of southern Xinjiang Uyghur Autonomous Region in 2021. Using a Space-Time Extra-Trees model, we assessed the four-year (2017–2020) average concentrations of particulate matter with aerodynamic diameter ≤1 µm (PM1), particulate matter with aerodynamic diameter ≤2.5 µm (PM2.5), particulate matter with aerodynamic diameter ≤10 µm (PM10), ozone (O3), sulfur dioxide (SO2), and carbon monoxide (CO), and then assigned these values to the participants. Generalized linear mixed models were employed to examine the relationships between air pollutants and the prevalence of MAFLD, with adjustment for multiple confounding factors. The odds ratios and 95% confidence intervals of MAFLD were 2.002 (1.826–2.195), 1.133 (1.108–1.157), 1.034 (1.027–1.040), 1.077 (1.023–1.134), 2.703 (2.322–3.146) and 1.033 (1.029–1.036) per 10 µg/m3 increase in the 4-year average PM1, PM2.5, PM10, O3, SO2 and CO exposures, respectively. The robustness of the findings was confirmed by a series of sensitivities. In summary, long-term exposure to ambient air pollutants was associated with increased odds of MAFLD, particularly in males and individuals with unhealthy lifestyles.

Radiomics using non-contrast CT to predict hemorrhagic transformation risk in stroke patients undergoing revascularization.

This study explores whether textural features from initial non-contrast CT scans of infarcted brain tissue are linked to hemorrhagic transformation susceptibility. Stroke patients undergoing thrombolysis or thrombectomy from Jan 2012 to Jan 2022 were analyzed retrospectively. Hemorrhagic transformation was defined using follow-up magnetic resonance imaging. A total of 94 radiomic features were extracted from the infarcted tissue on initial NCCT scans. Patients were divided into training and test sets (7:3 ratio). Two models were developed with fivefold cross-validation: one incorporating first-order and textural radiomic features, and another using only textural radiomic features. A clinical model was also constructed using logistic regression with clinical variables, and test set validation was performed. Among 362 patients, 218 had hemorrhagic transformations. The LightGBM model with all radiomics features had the best performance, with an area under the receiver operating characteristic curve (AUROC) of 0.986 (95% confidence interval [CI], 0.971-1.000) on the test dataset. The ExtraTrees model performed best when textural features were employed, with an AUROC of 0.845 (95% CI, 0.774-0.916). Minimum, maximum, and ten percentile values were significant predictors of hemorrhagic transformation. The clinical model showed an AUROC of 0.544 (95% CI, 0.431-0.658). The performance of the radiomics models was significantly better than that of the clinical model on the test dataset (p < 0.001). The radiomics model can predict hemorrhagic transformation using NCCT in stroke patients. Low Hounsfield unit was a strong predictor of hemorrhagic transformation, while textural features alone can predict hemorrhagic transformation. Using radiomic features extracted from initial non-contrast computed tomography, early prediction of hemorrhagic transformation has the potential to improve patient care and outcomes by aiding in personalized treatment decision-making and early identification of at-risk patients. • Predicting hemorrhagic transformation following thrombolysis in stroke is challenging since multiple factors are associated. • Radiomics features of infarcted tissue on initial non-contrast CT are associated with hemorrhagic transformation. • Textural features on non-contrast CT are associated with the frailty of the infarcted tissue.

Prediction of Global Horizontal Irradiance Using an Explainable Data Driven Machine Learning Algorithms

Estimating global horizontal irradiance (GHI) with a high level of accuracy and precision is very challenging due to the volatile climate parameters and location constraints. To overcome this challenge, several machine learning (ML)-based techniques such as Decision Trees (DT), Random Forest (RF), Extreme Gradient Boosting (XGB), and Extra Trees (ET) are implemented to forecast the GHI. The first stage of model development is to select the optimal subset of features by using the variance inflation factor feature selection method. In the second stage, the selected features are fed into the ML models and trained. The predictive performance of the ML models is improved the result of removal of insignificant input features. The predictive accuracy of the ML models is compared and evaluated by performance metrics such as mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R 2). Conclusively, after feature selection it is seen that the ET algorithm outperforms the others because of its lowest MAE and RMSE value of 3.01 and 1.748, respectively, as compared to the other models, indicating its relevancy, legitimacy, and viability for the estimation of GHI. The higher R 2 value of 0.99 obtained by the ET model indicates that it is best fitted with the dataset. Additionally, optimal shapely additive explanation values have been used as feature attributions for determining the magnitude and direction of the impact of each feature on the outcome.