Search Results (352)

We propose an unbiased restricted estimator that leverages prior information to enhance estimation efficiency for the linear regression model. The statistical properties of the proposed estimator are rigorously examined, highlighting its superiority over several existing methods. A simulation study is conducted to evaluate the performance of the estimators, and real-world data on total national research and development expenditures by country are analyzed to illustrate the findings. Both the simulation results and real-data analysis demonstrate that the proposed estimator consistently outperforms the alternatives considered in this study. Full article

Forest fires are the result of poor land management and climate change. Depending on the type of the affected eco-system, they can cause significant biodiversity losses. This study was conducted in the Amazonas department in Peru. Binary data obtained from the MODIS satellite on the occurrence of fires between 2010 and 2022 were used to build the risk models. To avoid multicollinearity, 12 variables that trigger fires were selected (Pearson ≤ 0.90) and grouped into four factors: (i) topographic, (ii) social, (iii) climatic, and (iv) biological. The program Rstudio and three types of machine learning were applied: MaxENT, Support Vector Machine (SVM), and Random Forest (RF). The results show that the RF model has the highest accuracy (AUC = 0.91), followed by MaxENT (AUC = 0.87) and SVM (AUC = 0.84). In the fire risk map elaborated with the RF model, 38.8% of the Amazonas region possesses a very low risk of fire occurrence, and 21.8% represents very high-risk level zones. This research will allow decision-makers to improve forest management in the Amazon region and to prioritize prospective management strategies such as the installation of water reservoirs in areas with a very high-risk level zone. In addition, it can support awareness-raising actions among inhabitants in the areas at greatest risk so that they will be prepared to mitigate and control risk and generate solutions in the event of forest fires occurring under different scenarios. Full article

Statistical methods are crucial for a wide range of analytical processes, from exploration and explanation to prediction and inference. Over the years, there has been a major shift towards machine learning and artificial intelligence techniques due to their powerful capability in learning the complex relationships between data. However, there is a disadvantage with these technologies in that mechanisms to explain the associations between a model’s input features and its output decision-making are far fewer than in statistics. This lack of transparency is among the major reasons that prevent machine learning from being more widely utilized in numerous application domains. Beyond inexplicability, the lack of mechanisms for effectively statistically assessing feature significance, such as parsimony or the complexity–performance tradeoff, further limits users’ control over machine learning models. With such motivation, we are proposing a target permutation process for determination of statistical feature importance in differentiable models and neural networks. Compared to methods in the current literature, the switch to target permutation allows for the assessment of all input features simultaneously and the test results are strengthened with a statistical p-value for each feature. In addition, our test does not require the assumption of independence among inputs, as is prevalent in other works. Lastly, we empirically show that our target permutation process can identify highly nonlinear associations between features and target while being resilient to multicollinearity. The features marked as insignificant can be removed with minimal impact, and can even result in improved predictive performance. Full article

The effectiveness of data-driven landslide susceptibility mapping relies on data integrity and advanced geospatial analysis; however, selecting the most suitable method and identifying key regional factors remains a challenging task. To address this, this study assessed the performance of six machine learning models, including Convolutional Neural Networks (CNNs), Random Forest (RF), Categorical Boosting (CatBoost), their CNN-based hybrid models (CNN+RF and CNN+CatBoost), and a Stacking Ensemble (SE) combining CNN, RF, and CatBoost in mapping landslide susceptibility along the Karakoram Highway in northern Pakistan. Twelve geospatial factors were examined, categorized into Topography/Geomorphology, Land Cover/Vegetation, Geology, Hydrology, and Anthropogenic Influence. A detailed landslide inventory of 272 occurrences was compiled to train the models. The proposed stacking ensemble and hybrid models improve landslide susceptibility modeling, with the stacking ensemble achieving an AUC of 0.91. Hybrid modeling enhances accuracy, with CNN–RF boosting RF’s AUC from 0.85 to 0.89 and CNN–CatBoost increasing CatBoost’s AUC from 0.87 to 0.90. Chi-square (χ²) values (9.8–21.2) and p-values (<0.005) confirm statistical significance across models. This study identifies approximately 20.70% of the area as from high to very high risk, with the SE model excelling in detecting high-risk zones. Key factors influencing landslide susceptibility showed slight variations across the models, while multicollinearity among variables remained minimal. The proposed modeling approach reduces uncertainties, enhances prediction accuracy, and supports decision-makers in implementing effective landslide mitigation strategies. Full article

Background/Objectives: Understanding the relationship between dietary patterns, nutrient intake, and chronic disease risk is critical for public health strategies. However, confounding from lifestyle and individual factors complicates the assessment of diet–disease associations. Emerging machine learning (ML) techniques offer novel approaches to clarifying the importance of multifactorial predictors. This study investigated the associations between animal-sourced and plant-based dietary patterns and Type 2 diabetes (T2D) history, accounting for diet–lifestyle patterns employing the XGBoost algorithm. Methods: Using data from the National Health and Nutrition Examination Survey (NHANES) from 2013 to 2016, individuals consuming animal-sourced foods (ASF) and plant-based foods (PBF) were propensity score-matched on key confounders, including age, gender, body mass index, energy intake, and physical activity levels. Predictors of T2D history were analyzed using the XGBoost classifier, with feature importance derived from Shapley plots. Lifestyle and dietary patterns derived from principal component analysis (PCA) were incorporated as predictors, and high multicollinearity among predictors was examined. Results: A total of 2746 respondents were included in the analysis. Among the top predictors of T2D were age, BMI, unhealthy lifestyle, and the ω6: ω3 fatty acid ratio. Higher intakes of protein from ASFs and fats from PBFs were associated with lower T2D risk. The XGBoost model achieved an accuracy of 83.4% and an AUROC of 68%. Conclusions: This study underscores the complex interactions between diet, lifestyle, and body composition in T2D risk. Machine learning techniques like XGBoost provide valuable insights into these multifactorial relationships by mitigating confounding and identifying key predictors. Future research should focus on prospective studies incorporating detailed nutrient analyses and ML approaches to refine prevention strategies and dietary recommendations for T2D. Full article

Old-growth forests are essential for maintaining biodiversity, as they are formed by the complexity of diverse forest structures, such as broad variations in tree height and diameter (DBH) and conditions of living and dead trees, leading to various ecological niches. However, many efforts of old-growth forest mapping from LiDAR have targeted only one specific forest structure (e.g., stand height, basal area, or stand density) by deriving information through a large number of LiDAR metrics. This study introduces a novel approach for identifying old-growth forests by optimizing a set of selected LiDAR standards and structural metrics. These metrics effectively capture the arrangement of multiple forest structures, such as canopy heterogeneity, multilayer canopy profile, and canopy openness. To determine the important LiDAR standard and structural metrics in identifying old-growth forests, multicollinearity analysis using the variance inflation factor (VIF) approach was applied to identify and remove metrics with high collinearity, followed by the random forest algorithm to rank which LiDAR standard and structural metrics are important in old-growth forest classification. The results demonstrate that the LiDAR structural metrics (i.e., advanced LiDAR metrics related to multiple canopy structures) are more important and effective in distinguishing old- and second-growth forests than LiDAR standard metrics (i.e., height- and density-based LiDAR metrics) using the European definition of a 150-year stand age threshold for old-growth forests. These structural metrics were then used as predictors for the final classification of old-growth forests, yielding an overall accuracy of 78%, with a true skill statistic (TSS) of 0.58 for the test dataset. This study demonstrates that using a few structural LiDAR metrics provides more information than a high number of standard LiDAR metrics, particularly for identifying old-growth forests in mixed temperate forests. The findings can aid forest and national park managers in developing a practical and efficient old-growth forest identification and monitoring method using LiDAR. Full article

Introduction: Colorectal cancer (CRC) is the third most commonly diagnosed cancer in the United States (U.S.). Mississippi has the highest rate of CRC incidence in the U.S. and has large populations of black and white individuals, allowing for studies of racial disparities. Methods: We conducted an ecological study using the county as the unit of analysis. CRC incidence data at the county level for black and white populations in Mississippi, covering the years 2003 to 2020, were retrieved from the Mississippi Cancer Registry. Age-adjusted incidence rate differences and their corresponding 95% confidence intervals (CIs) were then calculated for these groups. Getis–Ord Gi* hot and cold spot analysis of CRC incidence rate racial disparities was performed using ArcGIS Pro. We used global ordinary least square regression and geographically weighted regression (MGWR version 2.2) to identify factors associated with racial differences in CRC incidence rates. Results: Age-adjusted CRC incidence rate in the black population (median = 58.12/100,000 population) and in the white population (median = 46.44/100,000 population) varied by geographical area. Statistically significant racial differences in CRC incidence rates were identified in 28 counties, all of which showed higher incidence rates among the black population compared to the white population. No hot spots were detected, indicating that there were no spatial clusters of areas with pronounced racial disparities. As a post hoc analysis, after considering multicollinearity and a directed acyclic graph, a parsimonious multiple regression model showed an association (β = 0.93, 95% CI: 0.25, 1.62) indicating that a 1% increase in food insecurity was associated with a 0.93/100,000 differential increase in the black–white CRC incidence rate. Geographically weighted regression did not reveal any local patterns in this association. Conclusions: Black–white racial disparities in CRC incidence were found in 28 counties in Mississippi. The county-level percentage of food insecurity emerged as a possible predictor of the observed black–white racial disparities in CRC incidence rates. Individual-level studies are needed to clarify whether food insecurity is a driver of these disparities or a marker of systemic disadvantage in these counties. Full article

The purpose of this research is to examine the relationship between digital financial inclusion and economic growth in the SADC countries, while exploring the crucial moderating role of institutions. The digital financial inclusion (DFI) and institutional quality indices were constructed via Principal Component Analysis (PCA) to overcome the issue of multicollinearity. Using annual data from 2010 to 2023 and employing the system GMM technique, the findings of this study have persuasively supported the existence of a positive relationship between digital financial inclusion and economic growth in SADC countries, which signals that DFI is vital for boosting economic growth. The interaction term between the DFI index and institutional quality index also suggests a positive relationship, highlighting the need for the establishment of robust institutions and sound macroeconomic policies to ensure growth in the regional bloc. On the policy front, the findings indicate that efforts to expand digital financial inclusion in SADC countries should be complemented by institutional reforms aimed at improving governance, regulatory frameworks, rule of law, and legal protections. More specifically, this study suggests that SADC countries should focus on strengthening governance and regulatory frameworks to ensure transparency, security, and effective management of digital financial services. Improving legal protections, particularly around data security and consumer rights, is crucial to building trust in digital finance. Policymakers should also prioritize expanding digital infrastructure, especially in underserved areas, and addressing issues like limited technology access and digital literacy. Furthermore, fostering innovation in the fintech sector and implementing inclusive policies targeting marginalized groups will help drive wider adoption of digital financial services. By combining these reforms with institutional strengthening, SADC countries can create a conducive environment for sustainable economic growth through enhanced digital financial inclusion. Full article

This article presents an investigation of the use of machine learning methodologies for the prediction of surface roughness in milling operations, using sensor data as the primary source of information. The sensors, which included current transformers, a microphone, and displacement sensors, captured comprehensive machining signals at a frequency of 10 kHz. The signals were subjected to preprocessing using the Savitzky–Golay filter, with the objective of isolating relevant moments of active material machining and reducing noise. Two machine learning models, namely Elastic Net and neural networks, were employed for the prediction purposes. The Elastic Net model demonstrated effective handling of multicollinearity and reduction in the data dimensionality, while the neural networks, utilizing the ReLU activation function, exhibited the capacity to capture complex, nonlinear patterns. The models were evaluated using the coefficient of determination (R²), which yielded values of 0.94 for Elastic Net and 0.95 for neural networks, indicating a high degree of predictive accuracy. These findings illustrate the potential of machine learning to optimize manufacturing processes by facilitating precise predictions of surface roughness. Elastic Net demonstrated its utility in reducing dimensionality and selecting features, while neural networks proved effective in modeling complex data. Consequently, these methods exemplify the efficacy of integrating data-driven approaches with robust predictive models to improve the quality and efficiency of surface. Full article

Logistic regression models encounter challenges with correlated predictors and influential outliers. This study integrates robust estimators, including the Bianco–Yohai estimator (BY) and conditionally unbiased bounded influence estimator (CE), with the logistic Liu (LL), logistic ridge (LR), and logistic KL (KL) estimators. The resulting estimators (LL-BY, LL-CE, LR-BY, LR-CE, KL-BY, and KL-CE) are evaluated through simulations and real-life examples. KL-BY emerges as the preferred choice, displaying superior performance by reducing mean squared error (MSE) values and exhibiting robustness against multicollinearity and outliers. Adopting KL-BY can lead to stable and accurate predictions in logistic regression analysis. Full article

This study has made significant findings in A-share market data processing and portfolio management. Firstly, by adopting the Lasso method and CPCA framework, we effectively addressed the problem of multicollinearity among feature indicators, with the Lasso method demonstrating superior performance in handling this issue, thus providing a new method for financial data processing. Secondly, Deep Feedforward Neural Networks (DFN) exhibited exceptional performance in portfolio management, significantly outperforming other evaluated machine learning methods, and achieving high levels of out-of-sample performance and Sharpe ratios. Additionally, we consistently identified price changes, earnings per share, net assets per share, and excess returns as key factors influencing predictive signals. Finally, this study combined the Lasso method with DFN, providing a new perspective and methodological support for asset pricing measurement in the financial field. Full article

To promote sustainable development and global prosperity, policymakers collaborate on strategically harnessing natural resources and promoting renewable energy consumption to stimulate economic growth. This study examines the resource curse hypothesis across eight South Asian countries, Nepal, Sri Lanka, the Maldives, Bhutan, Pakistan, India, Afghanistan, and Bangladesh, from 1996 to 2022 using the ARDL model, multicollinearity analysis, unit root testing, and cointegration techniques. The findings reveal diverse effects in both the short and long runs. Natural resource rents have varying impacts on economic performance, experiencing negligible or even negative effects in the short term. In contrast, others show positive long-term relationships between natural resource exploitation and economic growth. The analysis of key economic factors, such as human capital, capital investment, energy consumption, and trade openness, shows that each influences economic performance in specific and measurable ways. This study highlights the significant role that natural resources play in shaping economic outcomes, which tends to negatively affect investment in many instances, underscoring the need for efficient resource management to avoid potential economic stagnation. This result may stem from the high upfront costs of renewable energy infrastructure, which outweigh short-term benefits, and the lack of supportive policies for renewable energy projects. This research confirms the presence of the resource curse in South Asian countries, stressing the need for efficient resource management strategies to prevent economic instability and mismanagement. Governments must implement policies that promote trade diversity and openness while fostering sustainable growth through improved resource governance. Full article

Methane emissions from dairy farms are a significant driver of climate change, yet their relationship with farm-specific practices remains poorly understood. This study employs Sentinel-5P satellite-derived methane column concentrations as a proxy to examine emission dynamics across 11 dairy farms in Eastern Canada, using data collected between January 2020 and December 2022. By integrating advanced analytics, we identified key drivers of methane concentrations, including herd genetics, feeding practices, and management strategies. Statistical tools such as Variance Inflation Factor (VIF) and Principal Component Analysis (PCA) addressed multicollinearity, stabilizing predictive models. Machine learning approaches—Random Forest and Neural Networks—revealed a strong negative correlation between methane concentrations and the Estimated Breeding Value (EBV) for protein percentage, demonstrating the potential of genetic selection for emissions mitigation. Our approach refined concentration estimates by integrating satellite data with localized atmospheric modeling, enhancing accuracy and spatial resolution. These findings highlight the transformative potential of combining satellite observations, machine learning, and farm-level characteristics to advance sustainable dairy farming. This research underscores the importance of targeted breeding programs and management strategies to optimize environmental and economic outcomes. Future work should expand datasets and apply inversion modeling for finer-scale emission quantification, advancing scalable solutions that balance productivity with ecological sustainability. Full article

High-accuracy gully erosion susceptibility maps play a crucial role in erosion vulnerability assessment and risk management. The principal purpose of the present research is to evaluate the predictive power of individual machine learning models such as random forest (RF), decision tree (DT), and support vector machine (SVM), and ensemble machine learning approaches such as stacking, voting, bagging, and boosting with k-fold cross validation resampling techniques for modeling gully erosion susceptibility in the Oued El Abid watershed in the Moroccan High Atlas. A dataset comprising 200 gully points, identified through field observations and high-resolution Google Earth imagery, was used, alongside 21 gully erosion conditioning factors selected based on their importance, information gain, and multi-collinearity analysis. The exploratory results indicate that all derived gully erosion susceptibility maps had a good accuracy for both individual and ensemble models. Based on the receiver operating characteristic (ROC), the RF and the SVM models had better predictive performances, with AUC = 0.82, than the DT model. However, ensemble models significantly outperformed individual models. Among the ensembles, the RF-DT-SVM stacking model achieved the highest predictive accuracy, with an AUC value of 0.86, highlighting its robustness and superior predictive capability. The prioritization results also confirmed the RF-DT-SVM ensemble model as the best. These findings highlight the superiority of ensemble learning models over individual ones and underscore their potential for application in similar geo-environmental contexts. Full article

The urban heat island (UHI) effect, marked by higher temperatures in urban areas compared to rural ones, is a key indicator of human-driven environmental changes. This study aims to identify the key morphological parameters that primarily contribute to the development of surface urban heat island intensity (SUHII) and investigates the relationship between SUHII and urban morphology using land surface temperature (LST) data from the Sentinel-3 satellite. The research focuses on Milan and Lecce, analyzing how urban geometry affects SUHII. Factors such as building height, aspect ratio, sky visibility, and surface cover are examined using approximately 1000 satellite images from 2022 and 2023. The study highlights seasonal and diurnal variations in SUHII, with particular emphasis on HW periods. Through multicollinearity and multiple regression analyses, the study identifies the main morphological drivers influencing SUHII in the two cities, specifically the Impervious Surface Fraction (ISF) and Mean Building Height (HM). Milan consistently exhibits higher SUHII, particularly during HWs, while Lecce experiences a negative SUHII, especially during the summer, due to lower urban density, more vegetation, and the low soil moisture around the urban area. Both cities show positive SUHII values at night, which are slightly elevated during HWs. The heat wave analysis reveals the areas most susceptible to overheating, typically characterized by high urban density, with ISF and HM values in some cases above the 90th percentile (0.8 and 13.0 m, respectively) compared to the overall distribution, particularly for Milan. The research emphasizes the importance of urban morphology in influencing SUHII, suggesting that detailed morphological analysis is crucial for developing climate adaptation and urban planning strategies to reduce urban overheating and improve urban resilience to climate change. Full article