In arid and semi-arid lands like Iran water is scarce, and not all the wastewater can be treated. Hence, groundwater remains the primary and the principal source of water supply for human consumption. Therefore, this study attempted to spatially assess the groundwater potential in an aquifer in a semi-arid region of Iran using geographic information systems (GIS)-based statistical modeling. To this end, 75 agricultural wells across the Marvdasht Plain were sampled, and the water samples’ electrical conductivity (EC) was measured. To model the groundwater quality, multiple linear regression (MLR) and principal component regression (PCR) coupled with elven environmental parameters (soil-topographical parameters) were employed. The results showed that that soil EC (SEC) with Beta = 0.78 was selected as the most influential factor affecting groundwater EC (GEC). CaCO₃ of soil samples and length-steepness (LS factor) were the second and third effective parameters. SEC with r = 0.89 and CaCO₃ with r = 0.79 and LS factor with r = 0.69 were also characterized for PC1. According to performance criteria, the MLR model with R² = 0.94, root mean square error (RMSE) = 450 µScm⁻¹ and mean error (ME) = 125 µScm⁻¹ provided better results in predicting the GEC. The GEC map indicated that 16% of the Marvdasht groundwater was not suitable for agriculture. It was concluded that GIS, combined with statistical methods, could predict groundwater quality in the semi-arid regions.

The application of artificial neural networks (ANNs) in agricultural sciences has proved its importance as a mathematical modeling technique for prediction and providing information about the direct effects (relative importance) of the inputs. Many methods have been examined to measure the relative importance (direct or maineffects) of inputs: however, they never report any details regarding the relative importance of interactions among inputs. Indeed, the relative importance explains the role of predictors to better understand the effect of each predictor separately. A total of seven ANN based methods were used in this study to estimate the relative importance (effects) of EC, pH, CaCO₃, OM, Silt, and Clay on Cation Exchange Capacity (CEC), six of them were used to estimate main effects only, while the seventh one was proposed to estimate both main and interactional effects. The generalized weights (GW) method expresses the importance of independent variables or predictors in ANN models. Nevertheless, it cannot provide any information about the interactions among them. The current study proposes a modification algorithm to the GW method to assess the relative importance of both direct and interactional effects of the selected inputs. The modified generalized weights (MGW) have been examined on an empirical dataset. The empirical dataset presents some chemical soil properties (EC, pH, CaCO₃, OM, Silt, and Clay as inputs, and CEC as output). This dataset was divided into three subsets because Silt, Clay, and Sand have multicollinearity and should not be examined together when modeled in ANN. The first subset consists of EC, pH, CaCO₃, OM, Silt, and CEC. Silt was replaced with Clay and Sand in the second and third subsets, respectively. All the methods performed well except for Garson’s algorithm and improved Garson’s algorithm. The results of the three subsets revealed that Silt, Clay, and Sand were the most effective components to CEC, followed by EC, pH, CaCO₃, OM. Most of the interactions were negative, which means that when one property increased, the effect of the other property on CEC decreased. In contrast, positive interactions mean that when one property increased, the effect of the other property on CEC increased. The MGW seems to be more efficient in providing indexes regarding the direct effects of inputs and their interactions.

Cardiovascular diseases (CVDs) remain the leading cause of mortality in the U.S. and exhibit pronounced geographic variation. Although prior studies have d

Spatial epidemiology, defined as the study of spatial patterns in disease burdens or health outcomes, aims to estimate disease risk or incidence by identifying geographical risk factors and populations at risk (Morrison et al., 2024). Research in spatial epidemiology relies on both conventional approaches and Machine- Learning (ML) algorithms to explore geographic patterns of diseases and identify influential factors (Pfeiffer & Stevens, 2015). Traditional spatial techniques, including spatial autocorrelation using global Moran’s I, Geary’s C (Amgalan et al., 2022), and Ripley’s K Function (Kan et al., 2022), Local Indicators of Spatial Association (LISA) (Sansuk et al., 2023), hotspot analysis by Getis-Ord Gi* (Lun et al., 2022), spatial lag models (Rey & Franklin, 2022), and Geographically Weighted Regression (GWR) (Kiani et al., 2024) are designed to explicitly incorporate the spatial structure of data into spatial modelling, often referred to as spatially aware models (Reich et al., 2021). Beyond these models, several other spatially aware approaches that have been widely applied in epidemiological studies include but are not limited to Bayesian spatial models that account for spatial uncertainty in disease mapping, such as Bayesian Hierarchical models, Conditional Autoregressive (CAR), and Besage, York, and Mollie’ (BYM) models (Louzada et al., 2021). Bayesian methods are statistically rigorous techniques that assume neighboring regions share similar values. Kulldorff’s Spatial Scan Statistic is another traditional spatial technique that uses a moving circular window to extract significant disease clusters (Tango, 2021). Moreover, geostatistical models such as Kriging and Inverse Distance Weighting (IDW) allow for continuous spatial interpolation of health data (Nayak et al., 2021).