Ensemble Model Research Articles

Web-based Land Suitability Analysis for Horticulture Crop (Kiwi) in Champawat districts, Uttrakhand using MCDM and Geospatial Analysis

Identifying the suitability of a horticulture fruit with a defined area is an active research area, it has been attempted through different approaches, surveying and geographic information system are the two most popular techniques utilized extensively for this activity across the globe. GIS has its proven advantage over the field surveys based approach and modern GIS tools offers automation of the entire work flow for this activity. This study presents a web-based land suitability analysis for kiwi fruit cultivation in the Champawat District of Uttarakhand, using AHP, TOPSIS, and Ensemble modeling in conjunction with Multi-Criteria Decision Making (MCDM) and Geospatial Analysis. Land suitability was evaluated across various classes based on nine criteria viz. elevation, chilling hours, land use/land cover (LULC), soil texture, soil depth, temperature, rainfall, slope, and aspect. The AHP method identified 26.09 hectares as highly suitable, 14,606.19 hectares as suitable, 31,899.77 hectares as moderately suitable, 8,807.99 hectares as less suitable, and 192.23 hectares as not suitable. TOPSIS results showed 4,294.32 hectares as highly suitable, 14,193.54 hectares as suitable, 11,782.52 hectares as moderately suitable, and 2,399.21 hectares as less suitable. The ensemble method indicated 24.15 hectares as highly suitable, 8,763.51 hectares as suitable, 13,785.13 hectares as moderately suitable, 2,472.13 hectares as less suitable, and 191.94 hectares as not suitable. Integrating these findings revealed that the ensemble approach closely aligns with AHP in identifying highly suitable and less suitable areas, while TOPSIS emphasizes moderate suitability. This synthesis provides valuable insights for land use planning for kiwi cultivation. Additionally, an online spatial decision support system (SDSS) was developed using the Django framework, offering a modular and secure environment for informed decision-making.

QSAR Regression Models for Predicting HMG-CoA Reductase Inhibition

Background/Objectives: HMG-CoA reductase is an enzyme that regulates the initial stage of cholesterol synthesis, and its inhibitors are widely used in the treatment of cardiovascular diseases. Methods: We have created a set of quantitative structure-activity relationship (QSAR) models for human HMG-CoA reductase inhibitors using nested cross-validation as the primary validation method. To develop the QSAR models, we employed various machine learning regression algorithms, feature selection methods, and fingerprints or descriptor datasets. Results: We built and evaluated a total of 300 models, selecting 21 that demonstrated good performance (coefficient of determination, R2 ≥ 0.70 or concordance correlation coefficient, CCC ≥ 0.85). Six of these top-performing models met both performance criteria and were used to construct five ensemble models. We identified the descriptors most important in explaining HMG-CoA inhibition for each of the six best-performing models. We used the top models to search through over 220,000 chemical compounds from a large database (ZINC 15) for potential new inhibitors. Only a small fraction (237 out of approximately 220,000 compounds) had reliable predictions with mean pIC50 values ≥ 8 (IC50 values ≤ 10 nM). Our svm-based ensemble model predicted IC50 values < 10 nM for roughly 0.08% of the screened compounds. We have also illustrated the potential applications of these QSAR models in understanding the cholesterol-lowering activities of herbal extracts, such as those reported for an extract prepared from the Iris × germanica rhizome. Conclusions: Our QSAR models can accurately predict human HMG-CoA reductase inhibitors, having the potential to accelerate the discovery of novel cholesterol-lowering agents and may also be applied to understand the mechanisms underlying the reported cholesterol-lowering activities of herbal extracts.

Integrated metaheuristic approaches for estimation of fracture porosity derived from fullbore formation micro-imager logs: Reaping the benefits of stand-alone and ensemble machine learning models

Fracture porosity is one of the most effective parameters for reservoir productivity and recovery efficiency. This study aims to predict and improve the accuracy of fracture porosity estimation through the application of advanced machine learning (ML) algorithms. A novel approach was introduced for the first time to estimate fracture porosity by reaping the benefits of petrophysical and fullbore formation micro-imager (FMI) data based on employing various stand-alone, ensemble, optimisation and multi-variable linear regression (MVLR) algorithms. This study proposes a ground-breaking two-step committee machine (CM) model. Petrophysical data containing compressional sonic-log travel time, deep resistivity, neutron porosity and bulk density (inputs), along with FMI-derived fracture porosity values (outputs), were employed. Nine stand-alone ML algorithms, including back-propagation neural network, Takagi and Sugeno fuzzy system, adaptive neuro-fuzzy inference system, decision tree, radial basis function, extreme gradient boosting, least-squares boosting, least squares support vector regression and k-nearest neighbours, were trained for initial estimation. To improve the efficacy of stand-alone algorithms, their outputs were combined in CM structures using optimisation algorithms. This integration was applied through five optimisation algorithms, including genetic algorithm, ant colony, particle swarm, covariance matrix adaptation evolution strategy (CMA-ES) and Coyote optimisation algorithm. Considering the lowest error, the CM with CMA-ES showed superior performance. Subsequently, MVLR was employed to improve the CMs further. Employing MVLR to combine the CMs yielded a 57.85% decline in mean squared error and a 4.502% improvement in the correlation coefficient compared to the stand-alone algorithms. The results of the benchmark analysis validated the efficacy of this approach.

Statistical Relationship between Satellite observed Land Surface Temperature and in-situ measured Surface Air Temperature over the Indian Region: An Exploratory Study

This paper presents an investigation of spatio-temporal estimation of daily minimum (Tmin) and maximum (Tmax) surface air temperature using satellite (INSAT and MODIS) derived Land surface temperature and in-situ observational data over the Indian region. Least Absolute Shrinkage and Selection Operator (LASSO) regression technique is used to identify influencing neighboring stations. To capture spatial and temporal variability of surface air temperature of a particular IMD station, Linear Regression and Auto-Regressive Integrated Moving Average, which is known as ARIMA, are used respectively. These models are statistically ensembled using stack generalization. Obtained station-by-station relationship is validated on an independent test data using Root Mean Squared Error (RMSE) to check the validity of the model under consideration. Results show ensembled model outperform (has lowest RMSE) the traditional methods for prediction of Tmin and Tmax with RMSE within a range of [1, 2] range for most of the regions and seasons.

Underestimation of meteorological drought intensity due to lengthening of the drought season with climate change

Abstract Meteorological drought may lead to water shortages, which has negative impacts on water-dependent sectors. Whilst there is a wealth of studies on changing drought intensity or frequency due to climate change, much less is known regarding potential shifts in the timing of drought. The purpose of this study is to analyse the timing of the drought season in the Netherlands and climatic changes therein, with a special focus on the onset of the drought season. Based on an analysis of meteorological observations in the Netherlands over the period 1965-2023, we conclude that the Dutch meteorological drought season has extended forward in time. On average, the drought season starts 16 days earlier in the period 1994-2023 compared to 1965-1993. This is mostly the result of an increase in potential evapotranspiration, while the amount of precipitation does not show a clear change at the start of the growing season. Using three climate model ensembles, we show that a forced climate change signal exists, but that natural variability also plays a role. Following this assessment of trends in meteorological variables, we analyse the consequences for the operational monitoring of meteorological drought. In the Netherlands, this is done by means of the `precipitation deficit’-indicator, based on a fixed-in-time starting point (1 April) of the drought season. The combination of this fixed starting point and the observed earlier onset of the drought season, means that in some years the indicator underestimates drought intensity, and that climatic trends are underestimated. We therefore advocate for an update of the operational drought indicator, such that meteorological drought occurring before 1 April will not be missed.

An ensemble convolutional neural network model for brain stroke prediction using brain computed tomography images

A stroke is a potentially fatal brain attack that causes an interruption in the blood supply to the brain. As a result, brain cells start to die due to a lack of oxygen and nutrients. After a stroke, every minute is critical. A million or more brain cells perish every minute during a stroke. The prompt identification of a stroke can prevent lasting brain damage or even save the patient’s life. Doctors advise computed tomography (CT) images of the brain for earlier stroke detection. If doctors delay CT diagnosis or may make erroneous diagnoses, this can be life-threatening. For that reason, an automatic diagnosis of stroke from a brain CT scan image will be beneficial for stroke patients. This study moderates three pre-trained convolutional neural network (CNN) models named Inceptionv3, MobileNetv2, and Xception by updating the top layer of those models using the transfer-learning technique based on CT images of the brain. A new ensemble convolutional neural network (ENSNET) model is proposed for automatic brain stroke prediction from brain CT scan images. ENSNET is the average of two improved CNN models named InceptionV3 and Xception. We have relied on the following metrics: accuracy, precision, recall, f1-score, confusion matrix, accuracy versus epoch, loss versus epoch, and the receiver operating characteristic (ROC) curve to assess performance matrices. The accuracy of the moderated Inceptionv3 is 97.48%, the moderated MobileNetv2 is 83.29%, and the moderated Xception is 96.11%. Nonetheless, the suggested ensemble model ENSNET performs better than the other models when it comes to the diagnosis of stroke from brain CT scans, providing 98.86% accuracy, 97.71% precision, 98.46% recall, 98.08% f1-score, and 98.74% area under the ROC curve(AUC). Therefore, the proposed model ENSNET can detect strokes from computed tomography images of the brain more successfully than other models.

A Fast and Cost-Effective Electronic Nose Model for Methanol Detection Using Ensemble Learning

Methanol, commonly used to cut costs in the production of counterfeit alcohol, is extremely harmful to human health, potentially leading to severe outcomes, including death. In this study, an electronic nose system was designed using 11 inexpensive gas sensors to detect the proportion of methanol in an alcohol mixture. A total of 168 odor samples were taken and analyzed from eight types of ethanol–methanol mixtures prepared at different concentrations. Only 4 features out of 264 were selected using the feature selection method based on feature importance. These four features were extracted from the data of MQ-3, MQ-4, and MQ-137 sensors, and the classification process was carried out using the data of these sensors. A Voting Classifier, an ensemble model, was used with Linear Discriminant Analysis, Support Vector Machines, and Extra Trees algorithms. The Voting Classifier achieved 85.88% classification accuracy before and 81.85% after feature selection. With its cost effectiveness, fast processing time, and practicality, the recommended system shows great potential for detecting methanol, which threatens human health in counterfeit drink production.

Integration of SPEI and machine learning for assessing the characteristics of drought in the middle ganga plain, an agro-climatic region of India.

Drought, as a natural and intricate climatic phenomenon, poses challenges with implications for both natural ecosystems and socioeconomic conditions. Evaluating the characteristics of drought is a significant endeavor aimed at mitigating its impact on society and individuals. This research paper explores the integration of the Standardized Precipitation Evapotranspiration Index (SPEI) and machine learning techniques for an assessment of drought characteristics in the Middle Ganga Plain, a crucial agro-climatic region in India. The study focuses on evaluating the frequency, intensity, magnitude, and recurrence interval of drought events. Various drought models, including Random Forest (RF), Artificial Neural Networks (ANN), and an ensemble model combining ANN and RF, were employed to analyze and predict drought patterns at different temporal scales (3-month, 6-month, and 12-month). The performance of these models was rigorously validated using key metrics such as precision, accuracy, proportion incorrectly classified, over-all area under the curve (AUC), mean absolute error (MAE), and root mean square error (RMSE). Furthermore, the research extends its application to delineating drought vulnerability zones by establishing demarcations for high and very high drought vulnerability areas for each model and temporal scale. Results indicate that the south-western part of the middle Ganga plain falls under the highly drought-vulnerable zone, which averagely covers 40% of the study region. The core and buffer regions of drought vulnerability have also been identified. The south-western part of the study area is identified as the core region of drought. Ground verification of the drought-vulnerable area has been done by using soil moisture meter. Validation metrics show that the ensemble model of ANN and RF exhibits the highest accuracy across all temporal scales. This research's findings can be applied to improve drought preparedness and water resource management in the Middle Ganga Plain. By identifying high-risk drought zones and utilizing accurate prediction models, policymakers and farmers can implement targeted mitigation strategies. This approach could enhance agricultural resilience, protect livelihoods, and optimize water allocation in this vital agro-climatic region.

Investigation of the Features Influencing the Accuracy of Wind Turbine Power Calculation at Short-Term Intervals

The accurate prediction of wind power generation, as well as the development of a digital twin of a wind turbine, require estimation of the power curve. Actual measurements of generated power, especially over short-term intervals, show that in many cases the power generated differs from the calculated power, which considers only the wind speed and the technical parameters of the wind turbine. Some of these measurements are erroneous, while others are influenced by additional factors affecting generation beyond wind speed alone. This study presents an investigation of the features influencing the accuracy of calculations of wind turbine power at short-term intervals. The open dataset of SCADA-system measurements from a real wind turbine is used. It is discovered that using ensemble machine learning models and additional features, including the actual power from the previous time step, enhances the accuracy of the wind power calculation. The root-mean-square error achieved is 113 kW, with the nominal capacity of the wind turbine under consideration being 3.6 MW. Consequently, the ratio of the root-mean-square error to the nominal capacity is 3%.

Investigating uncertainties in air quality models used in GMAP/SIJAQ 2021 field campaign: General performance of different models and ensemble results

The international field campaign, GMAP/SIJAQ 2021, was conducted in Korea from October 18th to November 25th to enhance the performance and validation of the Geostationary Environment Monitoring Spectrometer (GEMS) products algorithm and obtain a better understanding of the current air pollution status of the Korean Peninsula. Five chemical transport models (CTMs), including CMAQ, CMAQ-GIST, CAMx, WRF-Chem, and WRF GEOS-Chem, were utilized during the campaign to assist in organizing the observation plan and identifying changes in pollutant concentrations and their spatiotemporal distribution in Korea following the Korea–United States Air Quality (KORUS-AQ) 2016. In this study, we evaluated the forecasting performance, strengths, and limitations of these five CTMs and their ensemble in simulating air quality. Intensive measurement data and intercomparisons were employed to explain discrepancies between observed and simulated results. A comparison of the CTM ensemble results for PM2.5 and various gaseous pollutants between the current GMAP/SIJAQ 2021 and previous KORUS-AQ 2016 campaigns showed the R-value for the total mass PM2.5 concentration increased from 0.88 to 0.94. This improvement is related to CTM updates, including the emission inventory and better reproductions of the concentrations of gaseous species. However, the models consistently underestimated carbon monoxide (CO) concentrations, similar to the results from KORUS-AQ. This finding still suggests a further challenge that requires consideration of missing anthropogenic sources. The results of the ensemble model agreed well with the chemical composition of PM2.5 observed at the intensive monitoring station. However, for NO3− and NH4+, discrepancies were primarily due to inaccuracies in the meteorological inputs, such as precipitation, relative humidity (RH), and nighttime planetary boundary layer height (PBLH) in the CTMs. Hence, all models overestimated the concentration of elemental carbon (EC), therefore, it is necessary to revise EC emissions in the SIJAQv2 inventory, as these apply to unusual levels recorded in Seoul during the reference year of 2018.

Development of machine learning models for prediction of left ventricular systolic dysfunction in patients with isolated left ventricular dilation

Abstract Background/Introduction Isolated left ventricular dilation (ILVD), in the absence of concomitant LV systolic dysfunction, is a risk factor for new reduction in LV ejection fraction (LVEF) and adverse clinical outcomes. We assessed the performance of different machine learning (ML) models in predicting new reduction in LVEF in patients with ILVD. Methods A retrospective analysis of de-identified patient records in the Tempus Database with at least two transthoracic echocardiograms (TTE) conducted between January 2019 and December 2022 were retrospectively analyzed, with the follow up period ending in January 2023. Inclusion/exclusion criteria are shown in Figure 1A. Briefly, patients with ILVD were defined based on echocardiographic parameters as those with LVEF > 50% and LV end diastolic volume indexed (LVEDVi) > 62 mL/m2 for females and > 75 mL/m2 for males, without prior LVEF < 50% and/or grade 2/3 diastolic dysfunction. The endpoint was LVEF < 40% on future (≥90 days) TTE. Five hard voting ML models and one soft voting ensemble model were trained with data parsed from TTE reports (Figure 1B). Feature importance (SHAP) was calculated, and receiver operating characteristic (ROC) and Kaplan Meier curves evaluated the final model (Figure 1C-E). Results A total of 2,962 patients met the inclusion criteria. The median time from index to last TTE was 578 (IQR = 500) days. LVEF < 40% on follow up TTE was noted in 6% of patients (177 patients, median 478 days to endpoint, IQR = 535 days). Of the six models developed, the ensemble model had the highest ROC area under the curve (0.791, Figure 1D), precision (0.163), F1-score (0.266), and specificity (0.773). The ensemble model also had the lowest predicted prevalence (25.5%), although this was higher than the observed prevalence in the test set (5.7%). Patients predicted to develop LVEF < 40% had a 8.2x higher risk of developing LVEF < 40% within five years of index (p<0.001, Figure 1E). Feature importance showed that male gender, higher heart rate, lower LVEF at index TTE, higher LVEDVi, and lower tricuspid annular plane systolic excursion had the highest mean absolute impact on the model’s output probability (Figure 1C). Conclusion Using only data parsed from the index TTE, ensembling a model for the prediction of developing LVEF < 40% in patients with ILVD and without a history of LV diastolic or systolic dysfunction performed the best. Additional analyses including clinical variables such as comorbidities, symptoms, and medications may improve the model’s performance.

DaE2: Unmasking malicious URLs by leveraging diverse and efficient ensemble machine learning for online security

Over 5.44 billion people now use the Internet, making it a vital part of daily life, enabling communication, e-commerce, education, and more. However, this huge Internet connectivity also raises concerns about online privacy and security, particularly with the rise of malicious Uniform Resource Locators (URLs). Recently, conventional ensemble models have attracted attention due to their notable benefits of reducing the variance in models, enhancing predictive performance, improving prediction accuracy, and demonstrating high generalization potential. But, its application in addressing the challenge of malicious URLs is still an open problem. These URLs often hide behind static links in emails or web pages, posing a threat to individuals and organizations. Despite blacklisting services, many harmful sites evade detection due to inadequate scrutiny or recent creation. Hence, to improve URL detection, a Diverse and Efficient Ensemble (DaE2) machine learning algorithm was developed using four ensemble models, that is, AdaBoost, Bagging, Stacking, and Voting to classify URLs. After preprocessing, the experimental result shown that all models achieved over 80 % accuracy, with AdaBoost reaching 98.5 % and Stacking offering the fastest runtime. AdaBoost and Bagging also delivered strong performance, with F1 scores of 0.980 and 0.976, respectively.

Prediction of family inheritance in non-sarcomeric hypertrophic cardiomyopathy using clinical variables and 12-lead electrocardiogram artificial intelligence models

Abstract Background Hypertrophic cardiomyopathy (HCM) can be familial even in the absence of a causative sarcomeric variant, justifying periodical screening of first-degree relatives. However, indiscriminate screening may burden low-risk individuals unnecessarily. The prevalence and predictors of familial disease in non-sarcomeric HCM remain inadequately characterized. Purpose Our study aims to develop a tailored approach to family screening based on index cases’ clinical profiles. By leveraging artificial intelligence (AI) models, we aim to predict the likelihood of family disease using pertinent clinical variables and electrocardiograms (ECGs) from the proband. Methods From January 2008 to June 2022, all probands diagnosed with HCM at our Heart Institute genetics clinic were included if they had i) negative genetic testing for a (likely) pathogenic variant in a sarcomeric gene and ii) ≥1 first- or second-degree relative screened for HCM. The outcome was the presence of family disease defined as left ventricular wall thickness ≥13mm or verified autopsy report confirming HCM. Predictors of family disease were assessed at the first genetic clinic visit and considered age, sex, hypertension, and echocardiography metrics (maximal wall thickness, septal morphology, outflow tract obstruction at rest or Valsava). Using 12-lead ECGs and clinical variables, we used machine learning to train two distinct AI models. Model A utilized ECG signals only to train a ResNet model, while Model B used both ECGs and selected clinical variables (below 0.2 alpha level in logistic regressions) to train a random forest model to predict family disease. We performed evaluations at both ECG-level (one ECG used per patient) and patient-level (all available ECGs used in a patient) and we present individual and ensemble model performance. Results Among 356 families with sarcomere-negative HCM, 106 (30%) were familial. Probands were aged 53 ± 16 years, 30% female, 36% with a diagnosis of hypertension and 43% with peak outflow tract obstruction ≥30mmHg. In univariable logistic regressions, family disease was associated with age at diagnosis (OR 0.89 per 5-year, P=0.003), hypertension (OR 0.60, P=0.046) and obstruction (OR 0.52, P=0.013). The ensemble AI model, which integrated clinical and ECG data, achieved an AUC of 0.82 (95% CI: [0.795, 0.848]). It demonstrated a specificity of 0.891 (95% CI: [0.867, 0.913]) and a negative predictive value of 0.798 (95% CI: [0.772, 0.823]). When comparing ECG-level and patient-level data (1 to 10 ECGs per patient), the models consistently achieved AUC values between 0.78 and 0.81, indicating stable performance across various data subsets. Conclusion The prevalence of family inheritance in sarcomere-negative HCM can be accurately predicted using AI models that integrate clinical variables and 12-lead ECG data. This approach could enable tailored screening strategies based on the likelihood of family disease.

An ensemble machine learning approach for predicting heart failure risk in acute coronary syndrome patients: heartguardai

Abstract Background Heart failure (HF) constitutes a major post-event morbidity in patients experiencing Acute Coronary Syndrome (ACS), underscoring the imperative for precise risk stratification methodologies. Traditional prognostic models, although beneficial, are limited by their reliance on a narrow scope of variables, thereby overlooking the comprehensive analytical potential afforded by extensive clinical datasets. Aims To develop "HeartGuardAI," an ensemble machine learning model that predicts HF risk in ACS patients. By incorporating a wide array of variables including ejection fraction (EF), glomerular filtration rate (GFR), body mas index(BMI), male sex, this model seeks to offer a more practical tool for clinical decision-making. Methods We queried ACTION-ACS a large-scale multicenter database of 3335 patients at 10-year follow-up. We employed a feature selection process informed by Random Forest Classifier to identify the most predictive variables. An ensemble model combining the strengths of Random Forest and Gradient Boosting methodologies was then developed. The model's predictive performance was validated through Stratified K-Fold cross-validation, focusing on Area Under the Curve (AUC) and confidence intervals to assess its efficacy. The model's predictive performance was validated through Stratified K-Fold cross-validation, focusing on Area Under the Curve (AUC) and confidence intervals to assess its efficacy. Results Demonstrating a mean AUC of 0.76 ± 0.036, "HeartGuardAI" showed a high capability to predict HF risk in post-ACS patients and emphasizes the significance of incorporating multifactorial clinical assessments Conclusions HF risk can be predicted by a handy tool developed with advanced machine learning methods. Future studies are needed to prospectively validate this algorithm.Graphic 1Graphic 2

Discovery of Plasma Lipids as Potential Biomarkers Distinguishing Breast Cancer Patients from Healthy Controls.

The development of a sensitive and specific blood test for the early detection of breast cancer is crucial to improve screening and patient outcomes. Existing methods, such as mammography, have limitations, necessitating the exploration of alternative approaches, including circulating factors. Using 598 prospectively collected blood samples, a multivariate plasma-derived lipid biomarker signature was developed that can distinguish healthy control individuals from those with breast cancer. Liquid chromatography with high-resolution and tandem mass spectrometry (LC-MS/MS) was employed to identify lipids for both extracellular vesicle-derived and plasma-derived signatures. For each dataset, we identified a signature of 20 lipids using a robust, statistically rigorous feature selection algorithm based on random forest feature importance applied to cross-validated training samples. Using an ensemble of machine learning models, the plasma 20-lipid signature generated an area under the curve (AUC) of 0.95, sensitivity of 0.91, and specificity of 0.79. The results from this study indicate that lipids extracted from plasma can be used as target analytes in the development of assays to detect the presence of early-stage breast cancer.

Assessing the impacts of climate change on the vulnerability of coastal aquifers to the seawater intrusion

In most coastal areas, groundwater is affected by saltwater intrusion. Climate change effects, such as the sea-level rise and rainfall change are the climatic factors affecting the saltwater intrusion into groundwater. This study examines the vulnerability of the Abdan-Lamidan coastal aquifer to the intrusion of Persian Gulf saline waters under different hydrological conditions. First, the SIMCLIM model was used to calculate the regional sea-level rise of the Gulf under 24 future AOGCMs projections. The results showed that the increase in water level under the ensemble of AOGCM models and two RCP2.6 and RCP8.5 scenarios is 6.7 and 7.8 cm in 2050 compared to 2013. Then, the GALDIT vulnerability index was used to assess the vulnerability of the aquifer under the current condition and future scenarios. The results showed that parts of the aquifer that are currently at lower levels would be more vulnerable to sea-level rise in the future periods.

A new fusion neural network model and credit card fraud identification.

Credit card fraud identification is an important issue in risk prevention and control for banks and financial institutions. In order to establish an efficient credit card fraud identification model, this article studied the relevant factors that affect fraud identification. A credit card fraud identification model based on neural networks was constructed, and in-depth discussions and research were conducted. First, the layers of neural networks were deepened to improve the prediction accuracy of the model; second, this paper increase the hidden layer width of the neural network to improve the prediction accuracy of the model. This article proposes a new fusion neural network model by combining deep neural networks and wide neural networks, and applies the model to credit card fraud identification. The characteristic of this model is that the accuracy of prediction and F1 score are relatively high. Finally, use the random gradient descent method to train the model. On the test set, the proposed method has an accuracy of 96.44% and an F1 value of 96.17%, demonstrating good fraud recognition performance. After comparison, the method proposed in this paper is superior to machine learning models, ensemble learning models, and deep learning models.

Prediction of the Relative Resource Abundance of the Argentine Shortfin Squid Illex argentinus in the High Sea in the Southwest Atlantic Based on a Deep Learning Model.

To analyze the impact of the marine environment on the relative abundance of Illex argentinus (high and low categories) in the southwest Atlantic, this study collected logbook data from Chinese pelagic trawlers from December 2014 to June 2024, including vessel position data and oceanographic variables such as sea surface temperature, 50 m and 100 m water temperature, sea surface salinity, sea surface height, chlorophyll-a concentration, and mixed layer depth. Vessel positions were used to enhance the logbook data quality, allowing an analysis of the annual trends in the resource center of this squid at a spatial resolution of 0.1° × 0.1° and a temporal resolution of ten days. The findings showed that the resource center is primarily located around 42° S in the north and between 45° S and 47° S in the south, with a trend of northward movement during the study period. Additionally, we constructed two ensemble learning models based on decision trees-AdaBoost and PSO-RF-aiming to identify the most critical environmental factors that affect its resource abundance; we found that the optimal model was the PSO-RF model with max_depth of 5 and n_estimators of 46. The importance analysis revealed that sea surface temperature, mixed layer depth, sea surface height, sea surface salinity, and 50 m water temperature are critical environmental factors affecting this species' resources. Given that deep learning models generally have shorter running times and higher accuracy than other models, we developed a CNN-Attention model based on the five most important input factors. This model achieved an accuracy of 73.6% in forecasting this squid for 2024, predicting that the population would first appear near the Argentine exclusive economic zone around mid-December 2023 and gradually move east and south thereafter. The predictions of the model, validated through log data, maintained over 70% accuracy during most periods at a time scale of ten days. The successful construction of the resource abundance forecasting model and its accuracy improvements can help enterprises save fuel and time costs associated with blind searches for target species. Moreover, this research contributes to improving resource utilization efficiency and reducing fishing duration, thereby aiding in lowering carbon emissions from pelagic trawling activities, offering valuable insights for the sustainable development of this species' resources.

Advancing Algorithmic Adaptability in Hyperspectral Anomaly Detection with Stacking-Based Ensemble Learning

Anomaly detection in hyperspectral imaging is crucial for remote sensing, driving the development of numerous algorithms. However, systematic studies reveal a dichotomy where algorithms generally excel at either detecting anomalies in specific datasets or generalizing across heterogeneous datasets (i.e., lack adaptability). A key source of this dichotomy may center on the singular and like biases frequently employed by existing algorithms. Current research lacks experimentation into how integrating insights from diverse biases might counteract problems in singularly biased approaches. Addressing this gap, we propose stacking-based ensemble learning for hyperspectral anomaly detection (SELHAD). SELHAD introduces the integration of hyperspectral anomaly detection algorithms with diverse biases (e.g., Gaussian, density, partition) into a singular ensemble learning model and learns the factor to which each bias should contribute so anomaly detection performance is optimized. Additionally, it introduces bootstrapping strategies into hyperspectral anomaly detection algorithms to further increase robustness. We focused on five representative algorithms embodying common biases in hyperspectral anomaly detection and demonstrated how they result in the previously highlighted dichotomy. Subsequently, we demonstrated how SELHAD learns the interplay between these biases, enabling their collaborative utilization. In doing so, SELHAD transcends the limitations inherent in individual biases, thereby alleviating the dichotomy and advancing toward more adaptable solutions.

Asymmetric fluctuation of overlapping dinucleosome studied by cryoelectron microscopy and small-angle X-ray scattering.

Nucleosome remodelers modify the local structure of chromatin to release the region from nucleosome-mediated transcriptional suppression. Overlapping dinucleosomes (OLDNs) are nucleoprotein complexes formed around transcription start sites as a result of remodeling, and they consist of two nucleosome moieties: a histone octamer wrapped by DNA (octasome) and a histone hexamer wrapped by DNA (hexasome). While OLDN formation alters chromatin accessibility to proteins, the structural mechanism behind this process is poorly understood. Thus, this study investigated the characteristics of structural fluctuations in OLDNs. First, multiple structures of the OLDN were visualized through cryoelectron microscopy (cryoEM), providing an overview of the tilting motion of the hexasome relative to the octasome at the near-atomistic resolution. Second, small-angle X-ray scattering (SAXS) revealed the presence of OLDN conformations with a larger radius of gyration than cryoEM structures. A more complete description of OLDN fluctuation was proposed by SAXS-based ensemble modeling, which included possible transient structures. The ensemble model supported the tilting motion of the OLDN outlined by the cryoEM models, further suggesting the presence of more diverse conformations. The amplitude of the relative tilting motion of the hexasome was larger, and the nanoscale fluctuation in distance between the octasome and hexasome was also proposed. The cryoEM models were found to be mapped in the energetically stable region of the conformational distribution of the ensemble. Exhaustive complex modeling using all conformations that appeared in the structural ensemble suggested that conformational and motional asymmetries of the OLDN result in asymmetries in the accessibility of OLDN-binding proteins.