Multivariate Nonlinear Regression Model Research Articles

Age-adjusted Charlson comorbidity index predicts postoperative mortality in elderly patients with hip fracture: A prospective cohort.

This study aimed to evaluate the clinical association between the age-adjusted Charlson comorbidity index (aCCI) and postoperative mortality in elderly patients. Elderly patients with hip fractures were screened from January 2015 to September 2019. After demographic and clinical characteristics were collected, linear and non-linear multivariate Cox regression models were used to identify the association between the aCCI and mortality. All analyses were performed using EmpowerStats and R software. A total of 2,657 patients were included in the study, and the mean follow-up duration was of 38.97 months. The mean aCCI score was 4.24 ± 1.09, and 977 (34.14%) died of all-cause mortality. The fully-adjusted linear multivariate Cox regression models showed the aCCI to be associated with mortality [hazard ratio (HR) = 1.31, 95% confidence interval (CI):1.21-1.41, P < 0.0001]. Patients in Q2 showed greater mortality (HR = 1.60, 95% CI: 1.23-2.09; P = 0.0005) than those in Q1; patients in Q3 showed greater mortality (HR = 2.18, 95% CI: 1.66-2.87; P < 0.001) than those in Q1. In addition, the P-value for the trend also showed a linear association in the three models (P < 0.0001). In the sensitivity analysis, propensity score matching was used, and the results were stable. The mortality risk of hip fractures increased by 31% when the aCCI increased by one unit. aCCI score was shown to be a good predictor of three-year mortality following hip fracture. http://www.chictr.org.cn/showproj.aspx?proj=152919, identifier ChiCTR2200057323.

The Effect of Hematocrit on All-Cause Mortality in Geriatric Patients with Hip Fractures: A Prospective Cohort Study

The present study aimed to evaluate the association between hematocrit (HCT) levels and all-cause mortality in geriatric hip fractures. Older adult patients with hip fractures were screened between January 2015 and September 2019. The demographic and clinical characteristics of these patients were collected. Linear and nonlinear multivariate Cox regression models were used to identify the association between HCT levels and mortality. Analyses were performed using EmpowerStats and the R software. A total of 2589 patients were included in this study. The mean follow-up period was 38.94 months. Eight hundred and seventy-five (33.8%) patients died due to all-cause mortality. Linear multivariate Cox regression models showed that HCT level was associated with mortality (hazard ratio [HR] = 0.97, 95% confidence interval [CI]: 0.96-0.99, p = 0.0002) after adjusting for confounding factors. However, the linear association was unstable and nonlinearity was identified. A HCT level of 28% was the inflection point for prediction. A HCT level of <28% was associated with mortality (HR = 0.91, 95% CI: 0.87-0.95, p < 0.0001), whereas a HCT level > 28% was not a risk factor for mortality (HR = 0.99, 95% CI: 0.97-1.01, p = 0.3792). We found that the nonlinear association was very stable in the propensity score-matching sensitivity analysis. The HCT level was nonlinearly associated with mortality in geriatric hip fracture patients and could be considered a predictor of mortality in these patients. ChiCTR2200057323.

The association between hemoglobin at admission and mortality of older patients with hip fracture: a mean 3-year follow-up cohort study.

Hip fracture patients often have a large drop in hemoglobin (Hgb) concentration that is associated with the initial trauma. However, there is no study of a correlationbetween Hgb concentrationat admission and short-term mortality. Thus, weevaluated a possible linear and nonlinear association between Hgb and mortalityforolder patients with hip fracture. Consecutive older patients who had hip fractures were screened between January 2015 and September 2019. Demographic and clinical characteristics were collected. Linear and nonlinear multivariate Cox regression models were used to identify association between Hgb at admission and mortality. All analyses were performed with EmpowerStats and R software. Two thousand five hundred eighty-nine patients were included in the study. There were 849 men and 1740 women. The mean age was 79.6 ± 6.8years. The mean follow-up was 39.0months. Nine hundred seven (35.0%) patients died for all-cause reasons. The mean Hgb at admission was 11.07 ± 1.95g/dL. Linear multivariate Cox regression models showed Hgb at admission was associated with mortality ([Hazard Ratio] HR 0.91, 95% CI 0.87-0.95, P < 0.0001) after adjusting for confounding factors. However, the linear association was unstable, and nonlinearity was found between Hgb at admission and mortality. The Hgb concentration of 9.8g/dL was an inflection point. A Hgb at admission < 9.8g/dL was associated with mortality (HR 0.81, 95% CI 0.74-0.89, P < 0.0001), whereas > 9.8g/dL was not a risk factor for mortality (HR 0.98, 95% CI 0.92-1.04, P = 0.4730). The Hgb concentrationat admission was nonlinearly associated with mortality of older patients with hip fracture, and Hgb at admission < 9.8g/dL was a risk predictor of 3-year mortality. ChiCTR2200057323.

Evaluation of the Association between Low-Density Lipoprotein (LDL) and All-Cause Mortality in Geriatric Patients with Hip Fractures: A Prospective Cohort Study of 339 Patients

Many factors affect the prognosis of hip fractures in the elderly. Some studies have suggested a direct or indirect association among serum lipid levels, osteoporosis, and hip fracture risk. LDL levels were found to have a statistically significant nonlinear U-shaped relationship with hip fracture risk. However, the relationship between serum LDL levels and the prognosis of patients with hip fractures remains unclear. Therefore, in this study, we assessed the influence of serum LDL levels on patient mortality over a long-term follow-up period. Elderly patients with hip fractures were screened between January 2015 and September 2019, and their demographic and clinical characteristics were collected. Linear and nonlinear multivariate Cox regression models were used to identify the association between LDL levels and mortality. Analyses were performed using Empower Stats and R software. Overall, 339 patients with a mean follow-up period of 34.17 months were included in this study. Ninety-nine patients (29.20%) died due to all-cause mortality. Linear multivariate Cox regression models showed that LDL levels were associated with mortality (HR = 0.69, 95%CI: 0.53, 0.91, p = 0.0085) after adjusting for confounding factors. However, the linear association was unstable, and nonlinearity was identified. An LDL concentration of 2.31 mmol/L was defined as the inflection point for prediction. A LDL level < 2.31 mmol/L was associated with mortality (HR = 0.42, 95%CI: 0.25, 0.69, p = 0.0006), whereas LDL > 2.31 mmol/L was not a risk factor for mortality (HR = 1.06, 95%CI: 0.70, 1.63, p = 0.7722). The preoperative LDL level was nonlinearly associated with mortality in elderly patients with hip fractures, and the LDL level was a risk indicator of mortality. Furthermore, 2.31 mmol/L could be considered a predictor cut-off for risk.

Prediction Models of Shielding Effectiveness of Carbon Fibre Reinforced Cement-Based Composites against Electromagnetic Interference.

With the rapid development of communication technology as well as a rapid rise in the usage of electronic devices, a growth of concerns over unintentional electromagnetic interference emitted by these devices has been witnessed. Pioneer researchers have deeply studied the relationship between the shielding effectiveness and a few mixed design parameters for cementitious composites incoporating carbon fibres by conducting physical experiments. This paper, therefore, aims to develop and propose a series of prediction models for the shielding effectiveness of cementitious composites involving carbon fibres using frequency and mixed design parameters, such as the water-to-cement ratio, fibre content, sand-to-cement ratio and aspect ratio of the fibres. A multi-variable non-linear regression model and a backpropagation neural network (BPNN) model were developed to meet the different accuracy requirements as well as the complexity requirements. The results showed that the regression model reached an R2 of 0.88 with a root mean squared error (RMSE) of 2.3 dB for the testing set while the BPNN model had an R2 of 0.96 with an RMSE of 2.64 dB. Both models exhibited a sufficient prediction accuracy, and the results also supported that both the regression and the BPNN model are reasonable for such estimation.

Association between Neutrophil Levels on Admission and All-Cause Mortality in Geriatric Patients with Hip Fractures: A Prospective Cohort Study of 2,589 Patients.

To evaluate the association between neutrophil levels and all-cause mortality in geriatric hip fractures. Elderly patients with hip fractures were screened between January 2015 and September 2019. Demographic and clinical characteristics of the patients were collected. Linear and nonlinear multivariate Cox regression models were used to identify the association between neutrophil levels and mortality. Analyses were performed using Empower Stats and R software. A total of 2,589 patients were included in this study. The mean follow-up period was 38.95 months. During the study period, 875 (33.80%) patients died due to various causes. Linear multivariate Cox regression models showed that neutrophil levels were associated with mortality after adjusting for confounding factors, when neutrophil concentration increased by 1∗109/L, the mortality risk increased by 3% (HR = 1.03, 95% CI: 1.00-1.06, and P=0210). Neutrophil concentration was used as a categorical variable; we only found statistically significant differences when neutrophil levels were high (HR = 1.27, 95% CI:1.05-1.52, and P=0.0122). In addition, the results are stable in P for trend and propensity score matching sensitivity analysis. Neutrophil levels are associated with mortality in geriatric hip fractures and could be considered a predictor of death risk in the long-term. This study is registered with the Chinese Clinical Trial Registry (ChiCTR) as number ChiCTR2200057323.

Research on performance control of intercalated meltblown nonwovens based on neural network and grey relational analysis

In this paper, aiming at the performance of intercalated meltblown nonwoven materials, the relationship between its process parameters, structural variables and product performance is explored, and the product performance can be optimized by changing the process parameters through regression, planning and other methods. The results of grey correlation analysis showed that the intercalation rate had a significant effect on both structural variables and product properties. According to the results of the multivariate nonlinear regression model, the ideal maximum cleaning efficiency can reach 99.7%. Finally, in order to explore when the influencing factors such as acceptance distance, hot air speed, thickness, and compression rebound rate are limited, to achieve the goal of achieving the highest filtering efficiency and the smallest filtering resistance as possible, the machine learning method is used, and random forest is selected as the regression model, build samples and make predictions, and finally get the optimal value of the acceptance distance of 21 cm and the hot air speed of 1580 r/min.

Association of Kidney Function with Serum Levels of Cholesterol Absorption and Synthesis Markers: The CACHE Study CKD Analysis.

Serum levels of cholesterol absorption and synthesis markers are known to be associated with cardiovascular risk. Individuals with reduced kidney function or chronic kidney disease (CKD) are at an increased risk for cardiovascular disease. Hence, we examined the relationship between estimated glomerular filtration rate (eGFR) and serum markers of cholesterol absorption and synthesis. The CACHE (Cholesterol Absorption and Cholesterol synthesis in High-risk patiEnts) Consortium, comprised of 13 research groups in Japan possessing data of lathosterol (Latho, synthesis marker) and campesterol (Campe, absorption marker) measured via gas chromatography, compiled the clinical data using the REDCap system. Among the 3597 records, data from 2944 individuals were utilized for five analyses including this CKD analysis. This study analyzed data from 2200 individuals including 522 hemodialysis patients; 42.3% were female, the median age was 58 years, and the median eGFR was 68.9 mL/min/1.73 m2. Latho, Campe, and Campe/Latho ratio were significantly different when compared across CKD stages. When the associations of eGFR with these markers were assessed with multivariable nonlinear regression models, Latho, Campe, and Campe/Latho ratio showed positive, inverse, and inverse associations with eGFR. These associations were significantly modified by sex, the presence/absence of diabetes mellitus, and the presence/absence of statin use. We showed that individuals with lower eGFR have lower cholesterol synthesis marker levels and higher cholesterol absorption marker levels in this large sample.

Approximation of the Discharge Coefficient of Radial Gates Using Metaheuristic Regression Approaches

Radial gates are widely used for agricultural water management, flood controlling, etc. The existence of methods for the calculation of the discharge coefficient (Cd) of such gates are complex and they are based on some assumptions. The development of new usable and simple models is needed for the prediction of Cd. This study investigates the viability of a metaheuristic regression method, the Gaussian Process (GP), for the determination of the discharge coefficient of radial gates. For this purpose, a total of 2536 experimental data were compiled that cover a wide range of all the effective parameters. The results of GP were compared with the Group Method of Data Handling (GMDH), Multivariate Adaptive Regression Splines (MARS), and linear and nonlinear regression models for predicting Cd of radial gates in both free-flow and submerged-flow conditions. The results revealed that the radial basis function-based GP model performed the best in free-flow condition with a Correlation Coefficient (CC) of 0.9413 and Root Mean Square Error (RMSE) of 0.0190 while the best accuracy was obtained from the Pearson VII kernel function-based GP model for the submerged flow condition with a CC of 0.9961 and RMSE of 0.0132.

Nonlinear regression multivariate model for first order resonant periodic orbits and error analysis

In this work, nonlinear multivariate regression models are developed under the structure of the circular restricted three–body problem (CRTBP) with and without perturbations for exterior and interior resonant periodic orbits of order one. The nonlinear multivariate regression can be used for simply symmetric retrograde periodic orbits, which are first–order resonant orbits. Three real systems are taken into consideration for the study namely, Saturn–Hyperion, Saturn–Titan and Earth–Moon with small, moderate and high mass ratio respectively. In the given three systems, the first primary is taken as an oblate spheroid. Also, coefficient of oblateness, Jacobi constant and mass ratio are taken as parameters. The nonlinear multivariate regression model is used to find the initial position of periodic orbit without constructing PSS for time saving and avoid complicated procedure. In this context, fourth-fifth degree polynomial obtained from regression gives the best possible predicted value for the initial condition of periodic orbit by assigning parameter value such as mass ratio, coefficient of the oblateness of first primary and Jacobi constant. Also, error analysis of these models for three real systems Saturn–Hyperion, Saturn–Titan and Earth–Moon is discussed in detail. From error analysis, it can be seen that these models give the best approximation for the initial position of the periodic orbits. • Nonlinear multivariate regression model of the first order resonant periodic orbits in restricted three–body problem. • Analysis of periodic orbit for the Saturn–Hyperion, Saturn–Titan and Earth–Moon systems using Poincaré surface section. • Regression model and error analysis by considering oblateness of first primary body in the restricted three–body problem. • Estimation of maximum and minimum errors of the regression models for the perturbed circular restricted three–body problem.

Experimental Study on the Optimization of Coal-Based Solid Waste Filling Slurry Ratio Based on the Response Surface Method.

The large production and low comprehensive utilization rate of solid waste from coal power base affects the efficient and coordinated development of regional resources and the ecological environment. In order to promote utilization of solid waste from coal power base, coal gangue, fly ash, and gasification slag are mixed as raw materials to prepare filling materials, and a study on the evolution law of the mechanical properties of coal-based solid waste filling body is systematically carried out. After clarifying the physical and chemical properties of the filling materials, the Box–Behnken experimental design method was used to study the effects of slurry mass fraction, coal gangue, fly ash, and gasification slag on the strength of the filling body based on the response surface-satisfaction function coupling theory. Furthermore, a multivariate nonlinear regression model was constructed for the strength of the filling body at different maintenance ages. Based on the analysis of variance (ANOVA) and the response surface function, the impact mechanism of influencing factors and their interaction on the strength of filler were revealed. The results show that the strength of the filler is affected by single factors and interactions between factors. The interaction of slurry mass fraction and gangue dosing has a significant effect on the strength of the filler in the early stage; the interaction of fly ash and gangue dosing has a significant effect on the strength of the filler in the middle stage; the interaction of slurry mass fraction and gasification slag dosing has a significant effect on the strength of the filler in the final stage. The mixed filling materials significantly affect the strength of the filler as the maintenance time is extended. The mixed filling materials are extensively interlaced with the hydration products, calcium alumina, and calcium silicate hydrate (C-S-H) gel, forming a stable three-dimensional spatial support system as the maintenance time increases. The best ratio to meet the requirements of mine filling slurry pipeline transportation and filling body strength was selected using the regression model and the proposed economic function of filling material.

Research on a multivariate non-linear regression model of dynamic mechanical properties for the alkali-activated slag mortar with rubber tire crumb

As the inappropriate disposal of used tires will cause various environmental problems, rubber tire crumb (RTC) can be processed to replace fine aggregate in mortar. The content and particle size of RTC in mortar are the main influencing factors for the strength of rubber concrete. Alkali-activated slag mortar (AASM) can be used to replace ordinary Portland cement, which can improve the early strength and environmental friendliness of engineering. For this reason, the separating Hopkinson pressure bar (SHPB) test system was used to explore the effect of multi-factor coupling on the dynamic compressive strength of AASM. Besides, the significance analysis and multivariate non-linear regression analysis were carried out on the experimental results to determine the significance of each factor to the dynamic compressive strength of mortar. On this basis, The multivariate non-linear regression model was established for AASM dynamic compressive strength. The microstructure characterization and Statistics of chord length distribution of pore of AASM specimens were performed to explore the mechanism of strength deterioration experienced by rubber aggregate mortar. According to the experimental results, the surface of RTC can be modified in high alkaline environment, and the dynamic compressive strength of mortar would decline gradually with the increase of aggregate replacement ratio. In the optimal test group, the dynamic compressive strength of AASM is higher than that of the control group.

Influence of protein intake on the changes in skeletal muscle mass after kidney transplantation

Skeletal muscle mass decreases in patients with chronic kidney disease, especially those on dialysis with end-stage kidney disease. On the other hand, the recovery of renal function due to successful kidney transplantation (KT) improves skeletal muscle mass loss. However, low protein intake may influence the changes in skeletal muscle mass after KT. The aim of the present study is to examine the association of the changes in skeletal muscle mass with protein intake in kidney transplant recipients (KTRs). A cohort study was conducted in KTRs and living-kidney donors (LKDs). Skeletal muscle mass index (SMI) was measured using bioelectrical impedance analysis before KT and at 1 month and 12 months after KT. Protein intake was calculated with 24-h urine urea nitrogen from the Maroni formula at 12 months after KT. To evaluate the association between protein intake and the changes in SMI during the first year after KT, we performed a multivariable regression analysis adjusted for covariates including age, sex, cumulative glucocorticoids, cumulative hospitalization, diabetes mellitus, and SMI before KT. In KTRs (n=64), the median SMI was 7.26kg/m2 before KT, which decreased to 7.01kg/m2 at 1 month after KT and increased to 7.55kg/m2 at 12 months after KT. In LKDs (n=17), the median SMI was 6.24kg/m2 before KT which increased to 6.40kg/m2 at 1 month after KT and further increased to 6.95kg/m2 at 12 months after KT. The changes in SMI during the 1-year period after KT exhibited a positive correlation with protein intake (p=0.015) after adjustment. The predicted value of protein intake in KTRs, whose values of SMI before KT and at 12 months after KT were the same, was 0.72g/kg ideal body weight (IBW)/day using the multivariable non-linear regression model. In KTRs, insufficient protein intake adversely affected the recovery from skeletal muscle mass loss after KT. Therefore, a protein intake of at least more than 0.72g/kg IBW/day, the predicted value obtained in the present study, might be recommended for KTRs suffering from skeletal muscle mass loss.

Experimental Study on Compaction Characteristics and Mechanical Behavior of Crushed Rock Masses

The compaction and consolidation process of a regenerated roof can be divided into two stages, namely, secondary fragmentation of crushed rock masses and pore structure adjustment. Therefore, the mechanical properties of the regenerated roof differs significantly from those of the primary roof. We performed confined compression tests of crushed mudstones and sandy mudstones using a self-developed compression testing apparatus. We studied the influence of the moisture content, compression ratio, particle size, and volume-based degradation on the compression and consolidation characteristics based on the microscopic morphology of the crushed rock masses. Further, the degrees of influence of these factors on the stability of the consolidated bodies were determined. Therefore, the influence mechanism of each factor on the compressive and shear strengths of the consolidated bodies was revealed. The results indicated that the compaction and consolidation of the crushed rock masses involved a constant adjustment of the contact and stress states of the rock masses. Under uniaxial compression and shearing, the consolidated bodies presented the features of ductile failure. The compression ratio and particle size were the most important influencing factors in the compressive and shear strengths of the consolidated bodies. Further, we carried out an orthogonal experiment and built a multivariate nonlinear regression model to quantitatively characterize the combined influence of these factors on the compaction characteristics of the crushed rock masses and the mechanical behavior of the consolidated bodies.

Optimization of the whole-waste binder containing molten iron desulfurization slag from Kambara Reactor for concrete production

Kambara Reactor desulfurization slag (KRDS) is an alkaline solid waste, which is not properly utilized. The potential applications of KRDS in ground granulated blast-furnace slag (GGBS)-based cementitious materials and cement-free concrete were investigated in this study. An orthogonal test was conducted through range analysis and a multivariate nonlinear regression model to examine the effect of three factors: the GGBS content, KRDS/flue gas desulfurization gypsum (FGDG) ratio, and water/binder (W/B) ratio on the compressive strength of mortar. Next, a cement-free concrete was prepared with different strength grades using the cementitious material at the optimal ratios of raw materials. X-ray diffraction (XRD), thermogravimetry-differential scanning calorimetry (TG-DSC), Fourier transform infrared spectroscopy (FTIR), and scanning electron microscopy-energy dispersive X-ray spectrometry (SEM-EDS) analysis were carried out to reveal the hydration mechanism of the binder. The factors affecting the strength of the mortar are in this order: GGBS > KRDS/FGDG > W/B. The increase in KRDS enhances the early strength of the binder, but shows an adverse effect on its late strength. The optimum contents of GGBS, KRDS, and FGDG in the binder are 55%, 33.7% and 11.3%, respectively, with a W/B of 0.3. High levels of GGBS lead to an inhibitory effect of KRDS/FGDG on the compressive strength of the concrete. In order to improve the utilization rate and strength of KRDS, the level of GGBS should not exceed 55%. The main hydration products of the binder are needle-like ettringite (AFt) and spherical C-(A)-S-H gels. KRDS dissolution provides a strong alkaline environment, which promotes the disintegration of the Si(Al)–O tetrahedron of GGBS and the formation of the hydration products. C30–C40 concrete exhibit a small 1-h slump loss and good fluidity, making them more suitable for ready mix pumped concrete, whereas C50 and C60 concrete with a large slump loss are more suitable for high-strength precast members. This study has a theoretical and practical significance for a comprehensive utilization of multi-solid waste, energy saving, and carbon-emission reduction.

Estimation of surface particulate matter (PM2.5 and PM10) mass concentration by multivariable linear and nonlinear models using remote sensing data and meteorological variables over Ahvaz, Iran

Increased dust and air pollution in the Middle East over the past two decades has caused many problems, including human health risks and environmental hazards. Concentrations of surface Particulate Matter (PM 2.5 and PM 10 ) have always been considered as important indicators in assessing air pollution. It is necessary to have models that accurately estimate and predict the concentration of particulate matter to measure and reduce air pollution. In order to estimate the surface PM 2.5 and PM 10 concentrations, nine different linear and nonlinear multivariable regression models were provided in this study. Aerosol Optical Depth (AOD) data along with several effective meteorological variables such as temperature, relative humidity, wind speed, wind direction, horizontal visibility, and K-index were used to estimate PM 2.5 and PM 10 concentration. AOD data obtained from Medium Resolution Imaging Spectroscopy (MODIS) and the meteorological parameters were used to generate several statistical models, including linear and nonlinear multivariable regression models during ten years over Ahvaz. The highest correlation was observed between the observed and estimated PM 2.5 and PM 10 concentrations in the nonlinear equation. The forecast accuracy of the PM 2.5 and PM 10 concentrations using the nonlinear equation was 70% and 64%, respectively, which are the best predictions in the nine models obtained in this study. • Daily and monthly PM 2.5 and PM 10 estimated using the MODIS AOD data and ground-based meteorological parameters. • Multivariable linear and nonlinear regression analyses were used to estimate the concentration of PM 10 and PM 2.5 over Ahvaz. • Multivariable nonlinear regression models performed well with the forecast accuracy of 70% and 64% for PM 2.5 and PM 10

Cloning data with unchanged estimates of estimable non-linear functions of parameters

Non-linear regression models occur in the fields of biology, banking, economics, and sociology, population and biological growth. The absolute growth, growth of humans, and most importantly, an economic variable is appropriately described by non-linear regression models. In this article, we present cloned datasets for bivariate and multivariate non-linear regression models with the same non-linear regression fit. The application of such cloned datasets is used for the confidentiality of sensitive data for publication purposes. In this article, we present cloned data sets which will yield the same fitted non-linear regression models.

PM2.5 Concentration Prediction Based on Pollutant Pattern Recognition Using PCA-clustering Method and CS Algorithm Optimized SVR

Environmental issues, particularly air pollution, are a matter of concern for people all around the world. PM2.5 levels that are too high harm people’s physical and mental health. For government air pollution control, more accurate PM2.5 concentration predictions are critical. In this paper, we explored the relationship between pollutants (PM10, SO2, NO2, O3, CO) and meteorological factors (atmospheric pressure, relative humidity, air temperature, wind speed, wind direction, cumulative precipitation) that affect the generation and transmission of PM2.5. To better predict the concentration of PM2.5, we innovatively combined principal component analysis (PCA) and clustering methods to extract pollutant variables and patterns as important PM2.5 concentration predictors of different models such as support vector regression (SVR), multivariate nonlinear regression (MNR), and artificial neural network (ANN). Compared to MNR and ANN models, SVR presented better prediction accuracy. Moreover, cuckoo search (CS), cross-validation (CV), and particle swarm optimization (PSO) algorithms were used to further optimize the parameters in the process of SVR. And to evaluate the above PM2.5 concentration prediction results, we introduced several evaluating indicators including root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and person correlation coefficient (R) between predicted and measured values. The obtained results confirmed that when the pollutant data was divided into three patterns, the best prediction accuracy was achieved by the CS-SVR model.

Nonlinear Multivariate Regression Algorithms for Improving Precision of Multisensor Potentiometry in Analysis of Spent Nuclear Fuel Reprocessing Solutions

Potentiometric multisensor systems were shown to be very promising tools for the quantification of numerous analytes in complex radioactive samples deriving from spent nuclear fuel reprocessing. Traditional multivariate calibration for these multisensor systems is performed with partial least squares regression—an intrinsically linear regression method that can provide suboptimal results when handling potentiometric signals from very complex multi-component samples. In this work, a thorough investigation was performed on the performance of a multisensor system in combination with non-linear multivariate regression models for the quantification of analytes in the PUREX (Plutonium–URanium EXtraction) process. The multisensor system was composed of 17 cross-sensitive potentiometric sensors with plasticized polymeric membranes containing different lipophilic ligands capable of heavy metals, lanthanides, and actinides binding. Regression algorithms such as support vector machines (SVM), random forest (RF), and kernel-regularized least squares (KRLS) were tested and compared to the traditional partial least squares (PLS) method in the simultaneous quantification of the following elements in aqueous phase samples of the PUREX process: U, La, Ce, Sm, Zr, Mo, Zn, Ru, Fe, Ca, Am, and Cm. It was shown that non-linear methods outperformed PLS for most of the analytes.

Predicting medical usage rate at mass gathering events in Belgium: development and validation of a nonlinear multivariable regression model

BackgroundEvery year, volunteers of the Belgian Red Cross provide onsite medical care at more than 8000 mass gathering events and other manifestations. Today standardized planning tools for optimal preventive medical resource use during these events are lacking. This study aimed to develop and validate a prediction model of patient presentation rate (PPR) and transfer to hospital rate (TTHR) at mass gatherings in Belgium.MethodsMore than 200,000 medical interventions from 2006 to 2018 were pooled in a database. We used a subset of 28 different mass gatherings (194 unique events) to develop a nonlinear prediction model. Using regression trees, we identified potential predictors for PPR and TTHR at these mass gatherings. The additional effect of ambient temperature was studied by linear regression analysis. Finally, we validated the prediction models using two other subsets of the database.ResultsThe regression tree for PPR consisted of 7 splits, with mass gathering category as the most important predictor variable. Other predictor variables were attendance, number of days, and age class. Ambient temperature was positively associated with PPR at outdoor events in summer. Calibration of the model revealed an R2 of 0.68 (95% confidence interval 0.60–0.75). For TTHR, the most determining predictor variables were mass gathering category and predicted PPR (R2 = 0.48). External validation indicated limited predictive value for other events (R2 = 0.02 for PPR; R2 = 0.03 for TTHR).ConclusionsOur nonlinear model performed well in predicting PPR at the events used to build the model on, but had poor predictive value for other mass gatherings. The mass gathering categories “outdoor music” and “sports event” warrant further splitting in subcategories, and variables such as attendance, temperature and resource deployment need to be better recorded in the future to optimize prediction of medical usage rates, and hence, of resources needed for onsite emergency medical care.