Improving groundwater nitrate concentration prediction using local ensemble of machine learning models.

Groundwater is one of the most important water resources around the world, which is increasingly exposed to contamination. As nitrate is a common pollutant of groundwater and has negative effects on human health, predicting its concentration is of particular importance. Ensemble machine learning (ML) algorithms have been widely employed for nitrate concentration prediction in groundwater. However, existing ensemble models often overlook spatial variation by combining ML models with conventional methods like averaging. The objective of this study is to enhance the spatial accuracy of groundwater nitrate concentration prediction by integrating the outputs of ML models using a local approach that accounts for spatial variation. Initially, three widely used ML models including random forest regression (RFR), k-nearest neighbor (KNN), and support vector regression (SVR) were employed to predict groundwater nitrate concentration of Qom aquifer in Iran. Subsequently, the output of these models were integrated using geographically weighted regression (GWR) as a local model. The findings demonstrated that the ensemble of ML models using GWR resulted in the highest performance (R2 = 0.75 and RMSE = 9.38 mg/l) compared to an ensemble model using averaging (R2 = 0.68 and RMSE = 10.56 mg/l), as well as individual models such as RFR (R2 = 0.70 and RMSE = 10.16 mg/l), SVR (R2 = 0.59 and RMSE = 11.95 mg/l), and KNN (R2 = 0.57 and RMSE = 12.19 mg/l). The resulting prediction map revealed that groundwater nitrate contamination is predominantly concentrated in urban areas located in the northwestern regions of the study area. The insights gained from this study have practical implications for managers, assisting them in preventing nitrate pollution in groundwater and formulating strategies to improve water quality.

Spatial analysis and geoclimatic factors associated with the incidence of acute lymphoblastic leukemia in Iran during 2006-2014: An environmental epidemiological study.

BACKGROUND AND OBJECTIVES: The present study aims to determine the cumulative incidence rate of acute lymphoblastic leukemia (ALL), the degree of spatial autocorrelation and clustering of ALL, the hotspot and coldspots of ALL and geoclimatic conditions affecting the incidence of ALL in Iran and to draw a comparison between global and local regression models. MATERIALS AND METHODS: In this ecological study, an exploratory-etiologic multiple-group method has been adopted to investigate all children under 15 years of age with ALL in Iran during 2006-2014. Data analysis was performed using Mann Whitney U, Pearson correlation coefficients (PCCs), Global Moran's I, Optimized hotspot analysis (OHSA), Global Poisson regression (GPR), Geographically Weighted Poisson Regression (GWPR) at a significant level of α = 0.05. RESULTS: The cumulative incidence rate of ALL was estimated at 21,315 per 100,000 Iranian children under 15 years of age. The value of Global Moran's I index was estimated 0.338 and significant (<0.001 P-value). Coldspots were observed in north and northwest of Iran and hotspots were identified in south, southwest and mid-east of Iran. In the present study, Max Temperature of Warmest Month (MTWM) and Direct Normal Irradiation (DNI) were risk factors and Precipitation of the Coldest Quarter (PCQ) and Altitude (AL) were protective factors in the incidence of ALL, even though the non-stationarity of local coefficients and local t-values was clear. GWPR, by capturing and applying spatial heterogeneity and spatial autocorrelation, had a greater performance and goodness of fit than GPR. DISCUSSION: ALL has created spatial clusters in Iran. The incidence of ALL is the result of synergistic interaction between environmental, infectious, geographical and genetic risk factors. It is recommended to use of local models in ecological studies.

Identification of high-risk and low-risk clusters and estimation of the relative risk of acute lymphoblastic leukemia in provinces of Iran during 2006-2014 period: A geo-epidemiological study.

BACKGROUND: The present study was conducted to determine the epidemiological status, identify high-risk and low-risk clusters, and estimate the relative risk (RR) of acute lymphoblastic leukemia (ALL) in provinces of Iran. MATERIALS AND METHODS: This is an ecological study carried out using an Exploratory Multiple-Group design on 3769 children under 15 years of age with ALL from 2006 to 2014. Data analysis was performed using Mann-Whitney U, Global Moran's I and Kuldorff's purely spatial scan statistic tests at a significance level of 0.05. RESULTS: The average annual incidence rate of ALL during 2006-2014 period was 2.25/100,000 children under 15 years of age. The most likely high-risk cluster with log-likelihood ratio (LLR) =327.47 is located in the southwestern part of Iran with a radius of 294.93 km and a centrality of 30.77 N and 50.83 E, which contained 1276 patients with a RR of 2.56. It includes Fars, Bushehr, Kohgiluyeh and Boyer-Ahmad, Khuzestan and Chahar Mahall and Bakhtiari provinces. On the other hand, the most likely low-risk cluster with 517 patients, and a RR 0.49 and LLR = 227.03 was identified in the northwestern part of Iran with a radius of 270.38 km and a centrality of 37.25 N and 49.49 E. It includes Zanjan, Qazvin, Gilan and East Azerbaijan, Ardabil, Alborz and Tehran provinces. CONCLUSION: High-risk clusters were observed in Southwestern, central, and eastern Iran, while low-risk clusters were identified in Northern and Western Iran.

Geographical Pathology of Acute Lymphoblastic Leukemia in Iran with Evaluation of Incidence Trends of This Disease Using Joinpoint Regression Analysis.

BACKGROUND: The aim of present study is to determine the spatial-temporal epidemiology of acute lymphoblastic leukemia (ALL) in Iranian children. METHODS: This ecological study was performed using an exploratory mixed design. The study population consists of 3769 children with ALL who were reported in the National Cancer Registry Program of Iran from 2006 to 2014. Data analysis was conducted using Mann-Whitney U, joinpoint regression analysis, Global Moran's I and Anselin Local Moran's I. RESULTS: The average annual incidence rate of ALL was 2.25 per 100000 children under 15 years of age during the study period, which was 1.37 times higher in males. The average annual percentage change (AAPC) of the disease was 7.1%, which is higher than that of developed countries. The incidence of ALL was higher in spring and summer and its peak incidence was at the age of 2-5. Spatial autocorrelation of the ALL was 0.358 and significant (P value <0.001). The high-risk cluster of ALL was identified in Fars and Kohgiluyeh and Boyer-Ahmad provinces and the low-risk cluster in Kermanshah, Zanjan and Kurdistan provinces. CONCLUSION: The incidence of ALL is on rise in Iranian children and appropriate healthcare services are required to prevent new cases of this disease in high-risk areas.

Temporal trend and spatial distribution of acute lymphoblastic leukemia in Iranian children during 2006-2014: a mixed ecological study.

OBJECTIVES: The present study investigated the spatiotemporal epidemiological status of acute lymphoblastic leukemia (ALL), the most common childhood cancer, in Iran. METHODS: Using an exploratory mixed design, this ecological study examined 3,769 under-15 children with ALL recorded in the National Cancer Registry of Iran during 2006-2014. Data were analyzed using the Mann-Whitney U test, the Getis-Ord general G (GOGG) index, optimized hot spot analysis, and Pearson correlation coefficients (PCC) at a significance level of 0.05. RESULTS: The average annual incidence of the disease was 2.25 per 100,000 under-15 children, and the cumulative incidence rate (CIR) was 21.31 per 100,000 under-15 children. Patients' mean age was 5.90 years (standard deviation, 3.68), and the peak incidence was observed among 2-year to 5-year-olds. No significant difference was found in mean age between boys and girls (p=0.261). The incidence of ALL was more common during spring and summer than in other seasons. The GOGG index was 0.039 and significant (p<0.001). Hot spots were identified in south, central, and eastern Iran and cold spots in the north and west of Iran. The PCC between the CIR and latitude was negative (r=-0.507; p=0.003) but that between the CIR and longitude was positive (r=0.347; p=0.055). CONCLUSIONS: The incidence of ALL in Iranian children was lower than that observed in developed countries, but showed an increasing trend. It can be argued that the incidence of ALL is due to synergistic interactions between environmental, infectious, geographical, and genetic risk factors.

An epidemiological survey on road traffic crashes in Iran: application of the two logistic regression models.

Risk factors of human-related traffic crashes are the most important and preventable challenges for community health due to their noteworthy burden in developing countries in particular. The present study aims to investigate the role of human risk factors of road traffic crashes in Iran. Through a cross-sectional study using the COM 114 data collection forms, the police records of almost 600,000 crashes occurred in 2010 are investigated. The binary logistic regression and proportional odds regression models are used. The odds ratio for each risk factor is calculated. These models are adjusted for known confounding factors including age, sex and driving time. The traffic crash reports of 537,688 men (90.8%) and 54,480 women (9.2%) are analysed. The mean age is 34.1 ± 14 years. Not maintaining eyes on the road (53.7%) and losing control of the vehicle (21.4%) are the main causes of drivers' deaths in traffic crashes within cities. Not maintaining eyes on the road is also the most frequent human risk factor for road traffic crashes out of cities. Sudden lane excursion (OR = 9.9, 95% CI: 8.2-11.9) and seat belt non-compliance (OR = 8.7, CI: 6.7-10.1), exceeding authorised speed (OR = 17.9, CI: 12.7-25.1) and exceeding safe speed (OR = 9.7, CI: 7.2-13.2) are the most significant human risk factors for traffic crashes in Iran. The high mortality rate of 39 people for every 100,000 population emphasises on the importance of traffic crashes in Iran. Considering the important role of human risk factors in traffic crashes, struggling efforts are required to control dangerous driving behaviours such as exceeding speed, illegal overtaking and not maintaining eyes on the road.