Relative Importance Of Variables Research Articles

Knowledge-Informed Data-Driven Modeling of Coupled Human-Built–Natural Systems: The Case of Hurricane-Induced Debris

Debris is one of the most challenging cascading effects posed by hurricane events, causing large financial and logistical burdens to coastal communities. Moreover, disaster debris can have cascading consequences on the safety and functionality of infrastructure, inhibit community recovery efforts, and lead to public health concerns. However, existing debris predictive models have shown an unsatisfactory performance, with errors up to 50% in debris estimates, and only consider a limited set of predictive variables. Given the importance of debris in coastal community resilience, this study leveraged a convergent research strategy to propose a knowledge-informed data-driven methodology by developing a comprehensive database that expands across human-built–natural systems to inform a probabilistic data-driven model of debris volume. A Gaussian process model was used to generate the debris volume model from a debris removal database for Hurricane Ike and the human-built–natural systems predictors. Moreover, different spatial resolutions (500, 250, and 125 m) were tested to analyze their effect on the model performance. Results showed that the low-resolution (500 m) and the intermediate-resolution (250 m) models have the best performance with a normalized root-mean squared error (RMSE) of 0.49 and 0.50, respectively. These two models were then used to explore the relative variable importance of the predictive variables in the model in order to get insights on the drivers of the debris process and propose more flexible lower-dimensionality models. The influence of different predictors and the trade-offs of resolution and model performance were also discussed before demonstrating application of the model for a synthetic storm in the Galveston, Texas, region. The proposed methodology and the probabilistic estimates of debris quantities are key to develop comprehensive risk estimates of storm impacts on coastal communities and can support informed decision making and mitigation planning strategies.

Review on the Degradation of Poly(lactic acid) during Melt Processing.

This review paper presents an overview of the state of the art on process-induced degradation of poly(lactic acid) (PLA) and the relative importance of different processing variables. The sensitivity of PLA to degradation, especially during melt processing, is considered a significant challenge as it may result in deterioration of its properties. The focus of this review is on degradation during melt processing techniques such as injection molding and extrusion, and therefore it does not deal with biodegradation. Firstly, the general processing and fundamental variables that determine the degradation are discussed. Secondly, the material properties (for example rheological, thermal, and mechanical) are presented that can be used to monitor and quantify the degradation. Thirdly, the effects of different processing variables on the extent of degradation are reviewed. Fourthly, additives are discussed for melt stabilization of PLA. Although current literature reports the degradation reactions and clearly indicates the effect of degradation on PLA's properties, there are still knowledge gaps in how to select and predict the processing conditions that minimize process-induced degradation to save raw materials and time during production.

Physical Activity, but Not Glycaemic Load, Is Associated with Lower Real-Time Glycaemic Control in Free-Living Women with Gestational Diabetes Mellitus.

Maintaining blood glucose within the target range is the primary treatment goal for women with gestational diabetes mellitus (GDM). Foods with low glycaemic loads are recommended in clinical practice; however, the relative importance of other key lifestyle variables is unexplored. This pilot study explored the associations of glycaemic load, carbohydrates and physical activity parameters on blood glucose concentrations in free-living women with GDM. Twenty-nine women (28-30 weeks gestation, 34 ± 4 years) with GDM were enrolled. Continuous glucose monitoring, physical activity (ActivPAL inclinometer) and dietary intake and dietary quality were measured concurrently for 3 days. Pearson correlation analyses determined the association between glucose levels and lifestyle variables. Despite all receiving the same nutrition education, only 55% of women were following a low glycaemic load diet with a large range of carbohydrate intakes (97-267 g/day). However, the glycaemic load did not correlate with 3-hr postprandial glucose (r2 = 0.021, p = 0.56) or 24-h glucose iAUC (r2 = 0.021, p = 0.58). A significant relationship between total stepping time and lower 24-h glucose iAUC (r2 = 0.308, p = 0.02) and nocturnal glucose (r2 = 0.224, p = 0.05) was found. In free-living women with diet-controlled GDM, more physical activity, i.e., steps accumulated across the day, may be a simple and effective strategy for improving maternal blood glucose concentrations.

An interpretable machine learning framework to understand bikeshare demand before and during the COVID-19 pandemic in New York City

ABSTRACT In recent years, bikesharing systems have become increasingly popular as affordable and sustainable micromobility solutions. Advanced mathematical models such as machine learning are required to generate good forecasts for bikeshare demand. To this end, this study proposes a machine learning modeling framework to estimate hourly demand in a large-scale bikesharing system. Two Extreme Gradient Boosting models were developed: one using data from before the COVID-19 pandemic (March 2019 to February 2020) and the other using data from during the pandemic (March 2020 to February 2021). Furthermore, a model interpretation framework based on SHapley Additive exPlanations was implemented. Based on the relative importance of the explanatory variables considered in this study, share of female users and hour of day were the two most important explanatory variables in both models. However, the month variable had higher importance in the pandemic model than in the pre-pandemic model.

Crop Insurance Premium Recommendation System Using Artificial Intelligence Techniques

Purpose: The objective of this study is to build a crop insurance premium recommender model which will be fair to both crop insurance policy holders and crop insurance service providers. Theoretical Framework: The Nonparametric Bayesian Model (modified) is the name of the proposed model suggested by Maulidi et al. (2021) and it consists of six variables which are regional risk, cultivation time period, land area, claim frequency, discount eligibility (local variable) and premium. Discount eligibility variable is introduced to encourage right farming practices among farmers. Design/methodology/approach: Descriptive research method is used in this study as it is used to accurately represent the characteristics of a group of items. The population for this study is 943 respondents. The entire dataset is used for in-depth and accurate analysis. Five Artificial Intelligence models (Machine Learning models) are proposed for crop insurance premium prediction and they are Ada Boost Regressor, Gradient Boosting Regressor, Extra Trees Regressor, Support Vector Regressor and K-Neighbors Regressor. Among them Gradient Boosting Regression model has given the highest accuracy. Thus, Gradient Boosting Regression model is the most suitable model to be recommended for crop insurance premium prediction. Findings and Suggestions: Regional risk, land area, claim frequency and cultivation time period is the order of independent variables from highest to least in terms of regression coefficient. This relative importance helps Non-Banking Financial Companies (NBFCs) to suggest farmers that they should concentrate most on the regional risk or chances of crop failure in a particular region in which they are doing agriculture and least on the cultivation time period of a crop or the season in which a crop is cultivated. Two suggestions for future researchers are to extend this research work to other parts of Tamil Nadu and to apply hybrid machine learning techniques to the proposed model. Practical Implication: Unlike the existing formula-based traditional method used for calculating crop insurance premium, artificial intelligence models (machine learning models) can automatically learn the changes that take place with respect to the nature of variables in the proposed model and improve its accuracy based on new data. Hence, the crop insurance premium suggested by the most accurate model among the artificial intelligence models used in this study will be fair to both NBFCs and farmers. Here, fair means moderate. On the other hand, the crop insurance premium suggested by the existing formula-based method may not be fair in the long term as they cannot automatically learn the changes that take place with respect to the nature of variables in the proposed model and improve. Originality/value: In this research article, the relative importance of independent variables in the proposed model is determined and it helps NBFCs to suggest farmers that they should concentrate most on the region they are doing agriculture and least on the cultivation time period of a crop. Additionally, a machine learning model which can automatically learn and improve itself is used and hence the crop insurance premium predicted by it will be fair. Finally, the entire population containing 943 respondents details is analysed.

Neglected tropical diseases risk correlates with poverty and early ecosystem destruction

BackgroundNeglected tropical diseases affect the most vulnerable populations and cause chronic and debilitating disorders. Socioeconomic vulnerability is a well-known and important determinant of neglected tropical diseases. For example, poverty and sanitation could influence parasite transmission. Nevertheless, the quantitative impact of socioeconomic conditions on disease transmission risk remains poorly explored.MethodsThis study investigated the role of socioeconomic variables in the predictive capacity of risk models of neglected tropical zoonoses using a decade of epidemiological data (2007–2018) from Brazil. Vector-borne diseases investigated in this study included dengue, malaria, Chagas disease, leishmaniasis, and Brazilian spotted fever, while directly-transmitted zoonotic diseases included schistosomiasis, leptospirosis, and hantaviruses. Environmental and socioeconomic predictors were combined with infectious disease data to build environmental and socioenvironmental sets of ecological niche models and their performances were compared.ResultsSocioeconomic variables were found to be as important as environmental variables in influencing the estimated likelihood of disease transmission across large spatial scales. The combination of socioeconomic and environmental variables improved overall model accuracy (or predictive power) by 10% on average (P < 0.01), reaching a maximum of 18% in the case of dengue fever. Gross domestic product was the most important socioeconomic variable (37% relative variable importance, all individual models exhibited P < 0.00), showing a decreasing relationship with disease indicating poverty as a major factor for disease transmission. Loss of natural vegetation cover between 2008 and 2018 was the most important environmental variable (42% relative variable importance, P < 0.05) among environmental models, exhibiting a decreasing relationship with disease probability, showing that these diseases are especially prevalent in areas where natural ecosystem destruction is on its initial stages and lower when ecosystem destruction is on more advanced stages.ConclusionsDestruction of natural ecosystems coupled with low income explain macro-scale neglected tropical and zoonotic disease probability in Brazil. Addition of socioeconomic variables improves transmission risk forecasts on tandem with environmental variables. Our results highlight that to efficiently address neglected tropical diseases, public health strategies must target both reduction of poverty and cessation of destruction of natural forests and savannas.

Modeling impacts of climate change on the distribution of invasive Opuntia ficus-indica (L.) Mill. in Ethiopia: Implications on biodiversity conservation

The main danger to biological diversity is the introduction of exotic species. Opuntia ficus-indica (O. ficus-indica) is a dangerous invasive species that has seriously harmed Ethiopia's ecology and economy. To properly inform decision-making about the control of this invasive species, it is crucial to investigate the projected invasion dynamics of O. ficus-indica in the country under the current climate change scenarios. Thus, the objective of this research was to evaluate the current distribution and relative importance of environmental variables for O. ficus-indica distribution, map the habitat's future suitability under scenarios of climate change and assess how habitat change would affect the species' future expected suitability in Ethiopia. The SDM R program was used to perform species distribution modeling (SDM) using 311 georeferenced presence records along with climatic variables. Predictive models were developed as an agreement model from six modeling methodologies to investigate the climatic suitability of target species for the years 2050 and 2070 under two shared socio-economic pathway (SSP) scenarios (SSP2-4.5 and 5–8.5) in order to estimate the risks of climate change to the species. Under the current climatic scenario, only 9.26% (104939.3 km2) and 4.05% (45850.6 km2) of the country were moderately and highly suitable for species dispersion and invasion respectively. The remaining 86.69% (980648 km2) was suitable for the distribution and invasion of the species. In 2050, under the SSP2-4.5 and 5–8.5, the highly suitable range of O. ficus-indica is anticipated to expand by 2.30% and 1.76%, whereas the moderately suitable area is predicted to decrease by 1.66% and 2.69%, respectively. Under the SSP2-4.5 and 5–8.5 scenarios, the highly suitable region for the species is expected to grow by 1.47% and 0.65%, respectively, in 2070 compared to the current climatic conditions. This invasive species had already had a considerable negative influence on rangelands in a significant portion of the country with the current cover. Its continuing growth would exacerbate the issue, cause significant economic and environmental harm, and endanger the community's way of living. If preventive and efficient management methods are not taken seriously, the species will have considerable negative environmental impacts, which would be one of the biggest difficulties for pastoralism and their livelihoods.

An analysis of the relative variable importance to flood fatality using a machine learning approach

Vietnam is regularly and severely affected by flood events and there were nearly 14,000 dead people in 200 separate floods from 1989 to 2015. However, there have been limited studies specifically on flood-related mortality in Vietnam. This paper presents a longitudinal investigation of flood fatalities in Vietnam. More specifically, we use the available national disaster database and machine learning techniques to investigate theimportance of different attributes of flood damage to the attribute of flood fatalities. The results show that thehousing damage attribute significantly influences the fatality attribute, of which the weights are 0.45, 0.62, and 0.36 for the random forest, boosting, and multiple linear regression techniques, respectively. Thus, it is recommended that the proper policy prioritize housing improvements, establish evacuation plans, and developa strategy for temporary flood shelters in flood-prone areas. Understanding how various components of flooddamage are more likely to lead to fatalities analyzed in this study is critical for developing risk reductionstrategies.

Disparity in the relative roles of biotic and abiotic drivers on tree mortality between warm-temperate and temperate forests in China

Tree mortality is a key component of carbon cycling and forest dynamics. However, the relative contributions of different mechanisms to tree mortality remain poorly understood, particularly in forests with different climates. To clarify the relative contributions of certain factors on tree mortality at various organizational levels, we censused the mortality of tree species in 4-ha Guandi Mountain (GDS) plot with a warm-temperate climate and in 25-ha Changbai Mountain (CBS) plot with a temperate climate in North and Northeast China, respectively. The generalized linear mixed model (GLMM) was implemented to examine the relative importance of biotic and abiotic factors on tree mortality. Of the variables tested, tree size was typically the most important factor in decreasing tree mortality at the community and guild levels in both GDS and CBS, and its negative effect was stronger in GDS. Neighbors’ variables had significantly positive effects on tree mortality in two plots. Differently, conspecific neighbors’ interactions had stronger facilitative effects on increasing tree mortality of community level, shade-tolerant, evergreen, and small trees (DBH < 10 cm) in GDS, while the heterospecific neighbors’ interactions played greater roles in facilitating tree mortality of community level, shade-tolerant, deciduous, small trees in CBS. Among the abiotic factors tested, convexity and soil nutrients had significant positive effects on increasing tree mortality of shade-tolerant, evergreen, and small trees in GDS. While the abiotic factors didn’t contribute significantly to tree mortality at the community and all guild levels in CBS. Accordingly, our study demonstrates that the relative importance of variables driving tree mortality varied dramatically among guilds and between warm-temperate and temperate forests. These results provide critical information for future studies assessing forest dynamics and offer insights contributing to forest management in different forests.

Microbial activity and habitat heterogeneity in surface soil in a high-mountain grassland landscape (Central Apennines, Italy)

A spatially intensive sampling of surface soil layers (81 stratified-randomized sampling points across a 14 km2 study area) was performed in a high-mountain grassland landscape (Central Apennines, Italy) in July 2017, in order to describe the heterogeneity of microbial habitats and measure their microbial enzymatic activity. Three different microbial habitat types were identified via Hierarchical Cluster Analysis on the basis of 14 (measured or remote-sensed) environmental variables (including chemical and physical soil properties, topographic and geomorphological features and vegetation cover): (i) sub-acidic, at high elevation, with high vegetation cover (of mostly acidophytic/mesophytic species) and high silt content, and the lowest clay content; (ii) sub-acidic, at low elevation, with a high vegetation cover and a very low stone cover, with the highest sand and the lowest silt content, and the highest available phosphorus; (iii) very shallow soil, mainly stone-covered, at high elevation, with the highest pH values, on steep slopes, with the highest content of organic matter and the highest water holding capacity. The third habitat showed the highest enzymatic activity (b-glucosidase, β-cellobiohydrolase and leucine-arylamidase) involved in C and N cycling, while the more acidic and deeper soils, typical of sinkhole or slope areas, favoured the acid phosphomonoesterase activity involved in P cycle thanks to the establishment of the rhizospheric activity of the herbaceous plant species. Moreover, we analysed the relative importance of environmental variables on the total enzyme activities via Boosted Regression Trees (BRT): the results highlighted - for the first time as far as we know - the importance of topographic factors such as slope inclination in predicting the microbial functional capacity in a mountain grassland ecosystem. We conclude that the different enzymatic activity patterns found in the three habitats suggest diverse microbial functions with respect to nutrient cycling, within a small landscape and a relatively homogeneous land-cover.

Ecosystem variables importance in the presence and abundance of a globally invasive fish

Changes in physical habitat that are associated with anthropogenic disturbances facilitate the establishment and expansion of non-native species in receiving environments. Here, we evaluated the relative importance of ecosystem variables for the presence and abundance of the invasive fish Poecilia reticulata in Brazil. We collected fish species and assessed environmental variables through an established physical habitat protocol in 220 stream sites located in southeastern and midwestern Brazil. A total of 14,816 P. reticulata individuals were collected in 43 stream sites, and 258 variables that describe the physical characteristics of streams were assessed, including measures of channel morphology, substrate size and type, habitat complexity and cover, riparian vegetation cover and structure, and human influence. Dimensionality reduction methods were employed to limit redundancy, resulting in a smaller set of the most relevant environmental variables. Subsequently, we used random forest models to assess the relative importance of these variables in determining the presence and abundance of P. reticulata. The presence of this invasive fish was primarily explained by human disturbance variables related to urbanization (total impact, pavement, artificial structure areal cover, riparian canopy cover, electrical conductivity, mean thalweg depth, and sand), whereas channel morphology (mean bank full height) and fish cover variables (natural fish cover, and aquatic macrophyte areal cover) were important predictors of its abundance. Identifying which ecosystem variables are favorable to the establishment of non-native species is an important step in preventing future biological invasions, as well as managing those that already occur.

Cumulative Effects of Comorbid Burden and Overactive Bladder Symptoms on Fall Risk Among Older Women Seeking Treatment for Urogynecologic Conditions.

Overactive bladder (OAB) syndrome and urinary incontinence, age, and comorbid burden are strong risk factors for falls in women. Less is known about their cumulative effects on fall risk in a urogynecologic population. The purpose of this study was to investigate the effects of coexisting OAB, older age, and comorbidities on risk of falling among treatment seeking women with pelvic floor disorders. We conducted a retrospective medical records review of 348 consecutive women presenting to a urogynecology clinic over 6 months. Fall risk was determined by the Centers for Disease Control and Prevention's, Stopping Elderly Accidents, Deaths, and Injuries screening tool. Clinical and sociodemographic measures were abstracted from the electronic medical record. Odds of screening positive for high fall risk based on different patient profiles were calculated. We then used a classification and regression tree analysis to determine the relative importance of the different variables on fall risk within the homogeneous subgroups. Of the 348 women (mean age, 58.7 ± 15.8 years) who completed the fall risk screen, 124 (36%) screened positive for increased fall risk. Overactive bladder symptoms increased the likelihood of a positive fall risk screen across all combinations of age and comorbid burden. The patient profile of ≥3 OAB symptoms, ≥4 comorbid conditions, and age 65 years or older increased the odds of screening positive for high fall risk more than 6-fold (odds ratio, 6.4; 95% confidence interval, 3.1-12.9). In the following order of importance, the combination of high comorbid burden, OAB, and older age identified approximately 3 in 4 patients (73.3%) at high risk of falling. The presence of 3 easily identifiable patient characteristics is strongly associated with a risk of falls in women seeking care for pelvic floor disorders.

In plateaus, land cover replaced climate as the vital environmental factor on the spatial composition of bumblebees with the altitude rising

Abstract Bumblebees show a series of advantages over other pollinating insects in offering pollination services for both the ecosystem and agricultural production, especially in high‐altitude areas. With climate warming, their diversity at high altitudes will be threatened. Therefore, more attention should be paid to the management of bumblebee resources at high altitudes to prevent their decline here. Discovering drivers responsible for the spatial variation of biodiversity is a benefit for developing appropriate management. Here, based on the bumblebee diversity on three plateau areas with an altitude gradient, the roles of different environmental factors, the climate, land cover, topography, habitat heterogeneity and soil, in the bumblebee composition were assessed by generalised dissimilarity modelling. And the relative importance of variables was evaluated by the impact on the magnitude of total biological change and the per cent change in deviance explained by models. According to the results, in the low‐altitude area, the climate played a vital role and the most important variable was the climate variable. Then, with the altitude rising, the effect of land cover continually increased. In the medium‐altitude area, the effects of different factors were detected to be similar, and the most important variables were a climate variable and a land cover variable. In the high‐altitude area, land cover finally replaced the climate as the most important factor, and one land cover variable was assessed as the most important variable. Our study revealed that the relative importance of different environmental factors on the biodiversity of bumblebees can differ as the altitude rises.

Spatial patterns and determinants of nitrogen composition in the trunk xylem sap of tree species from tropical to temperate forests

Trunk xylem sap nitrogen (N) composition can provide key information for interpreting tree N transport strategies in a changing environment. However, there remains poorly explored to properly depict the spatial patterns and determinants of the xylem sap N composition in different tree species across climate zones. Here, we analyzed the trunk xylem sap N of dominant tree species from tropical to temperate forests in eastern China. The random forest (RF) approach was used to quantify the relative importance of biotic and abiotic variables in driving the spatial patterns of the trunk xylem sap N composition. The concentration ranges of total N (TN), ammonium (NH4+) and nitrate (NO3−) in the trunk xylem sap were 0.37–37.23, 0.01–2.60 and 0.01–0.49 mM, respectively. Dissolved organic nitrogen (DON) was the main component of TN in the xylem sap of forest trees, accounting for 83.4 ± 3.6%, followed by NH4+ (11.8 ± 3.2%) and NO3− (4.8 ± 1.3%). In addition, the TN and NH4+ concentrations increased from the tropical rainforest and then decreased to the northern temperate forest, with a peak in the subtropical deciduous forest in the middle-latitude climatic transition zone. The NO3− concentration in the temperate forest was higher than that in other forests, regardless of coniferous and broad-leaf forests. Tree species was the key factor affecting the spatial TN patterns, whereas soil pH primarily and conversely determined the spatial NH4+ and NO3− patterns. Together, tree characteristics and soil properties were the most important factors affecting the spatial patterns of xylem sap N composition. These findings provide new insights into N cycling within forest trees and have implications for better understanding the adaptation to global changes in forest ecosystems.

Demographic and behavioural factors affecting public support for pedestrianisation in city centres: The case of Edinburgh, UK

This paper provides an integrated analytical framework to investigate the demographic and behavioural factors that significantly influence public support for pedestrianisation. Pedestrianisation is often introduced by local authorities with the intention of improving air quality, the walkability of streets, road safety and opportunities for the local economy, however, issues remain regarding how accessible pedestrianised areas are for individuals who have conditions that limit their mobility. Using data from a survey, conducted during 2020 in Edinburgh (UK), public perceptions towards pedestrianisation were investigated through statistical testing and the development of random forest and ordered probit models. The random forest approach can help identify the relative importance of explanatory variables, whereas the ordered probit models can unveil the demographic and behavioural determinants of public support. To account for the potential effect of unobserved heterogeneity within respondents’ perceptions, random parameters were also considered in the ordered probit modelling framework. Initial results showed that residents are generally supportive of most issues surrounding pedestrianisation. Random parameters ordered probit modelling identified mode of travel and trip frequency as significant factors affecting key aspects of public support, such that active travellers were significantly more likely to support pedestrianisation, while those who rarely visit Edinburgh city centre were more likely to oppose pedestrianisation. Overall, a variety of independent analyses and modelling approaches suggest common influences on opinion, including behavioural patterns relating to transport modal choice and trip frequency, while disability was also found to have considerable effect on support as a fixed and random parameter. The statistical models are evaluated in terms of goodness-of-fit measures, before policy implications are discussed.

Air pollution as a substantial threat to the improvement of agricultural total factor productivity: Global evidence

ObjectiveThis study aims to provide empirical evidence about whether and to what extent air pollution affects the global agricultural total factor productivity (TFP). MethodsThe research sample covers 146 countries all over the world during 2010–2019. Two-way fixed effects panel regression models are used to estimate air pollution’s impacts. A random forest analysis is conducted to assess the relative importance of independent variables. ResultsThe results show that, on average, a 1% increase in fine particulate matter (PM2.5) and tropospheric ozone (O3) concentration would cause the agricultural TFP to decline by 0.104% and 0.207%, respectively. Air pollution’s adverse impact widely exists in various countries with different development levels, pollution degrees, and industrial structures. This study also finds that temperature has a moderating effect on the relationship between PM2.5 and agricultural TFP. PM2.5 pollution’s detrimental impact is weaker (stronger) in a warmer (cooler) climate. In addition, the random forest analysis confirms that air pollution is among the most crucial predictors of agricultural productivity. ConclusionsAir pollution is a substantial threat to the improvement of global agricultural TFP. Worldwide actions should be taken to ameliorate air quality, for the sake of agricultural sustainability and global food security.

Acetaminophen adsorption to spherical carbons hydrothermally synthesized from sucrose: experimental, molecular, and mathematical modeling studies.

Acetaminophen (AAP) is an analgesic and non-steroidal anti-inflammatory drug and a micropollutant that has been detected in waterbodies worldwide. Here, we explore the characteristics of AAP adsorption onto spherical carbons (SCs) hydrothermally synthesized from pure sucrose as a carbon source. In one-factor-at-a-time experiments, the adsorption capacity of AAP remained relatively constant between pH 2 and 10 but became negligible at pH 12. The Raman, FTIR, and XPS spectra illustrate that hydrogen bonding, π-π interactions, and n-π* interactions could contribute to the AAP adsorption onto the SCs. CHEM3D modeling was used to explore hydrogen-bond formation, π-π interactions, n-π* interactions, and electrostatic repulsion between AAP and the SCs. In view of the pHpzc of the SCs (3.1) and the pKa of AAP (10.96), electrostatic repulsion could occur between negatively charged SCs and anionic AAP above pH 10. In consideration of the average pore diameter of the SCs (1.89nm) and the AAP molecular size (8.94Å × 7.95Å × 4.93Å), a pore-filling mechanism could contribute to the adsorption. A pseudo-second-order model was best fitted to the kinetic data (equilibrium time = 6h), whereas the Liu isotherm was most suitable for the equilibrium data (maximum adsorption capacity = 92.0mg/g). Adsorption of AAP to the SCs was exothermic at 10-40°C. The SCs were regenerated and reused for AAP adsorption using a methanol. Multiple-factor-at-once (MFAO) experiments (input variables: pH, temperature, adsorbent dosage, and initial AAP concentration; output: AAP adsorption capacity) were used to develop response surface methodology (RSM, quartic regression) and artificial neural network (ANN, topology 4:11:9:1) models. Analyses using additional MFAO experimental data reveal that the predictive ability of the ANN model (R2 = 0.890) was better than that of the RSM model (R2 = 0.764). Based on the weight values of the ANN model, the relative importance of the input variables on the output was quantified in the order of initial AAP concentration (100%) > adsorbent dosage (92.3%) > temperature (77.6%) > pH (43.6%).

Exploring the Relative Importance and Interactive Impacts of Explanatory Variables of the Built Environment on Ride-Hailing Ridership by Using the Optimal Parameter-Based Geographical Detector (OPGD) Model

The impact of the built environment on the ridership of ride-hailing results depends on the spatial grid scale. The existing research on the demand model of ride-hailing ignores the modifiable areal unit problem (MAUP). Taking Chengdu as an example, and taking the density of pick-ups and drop-offs as dependent variables, 12 explanatory variables were selected as independent variables according to the “5D” built environment theory. The nugget–sill ratio (NSR) method and optimal parameter-based geographical detector (OPGD) model were used to determine the optimal grid scale for the aggregation of the built environment variables and the ridership of ride-hailing. Based on the optimal grid scale, the optimal data discretization method of the explanatory variables was determined by comparing the results of the geographic detector under different discretization methods (such as the natural break method, k-means clustering method, equidistant method, and quantile method); we utilized the geographic detector model to explore the relative importance and the interactive impacts of the explanatory variables on the ridership of ride-hailing under the optimal grid scale and optimal data discretization method. The results indicated that: (1) the suggested grid scale for the aggregation of the built environment and ride-hailing ridership in Chengdu is 1100 m; (2) the optimal data discretization method is the quantile method; (3) the floor area ratio (FAR), distance from the nearest subway station, and residential POI (point of interest) density resulted in a relatively high importance of the explanatory variable that affects the ridership of ride-hailing; and (4) the interactions of the diversity index of mixed land use ∩ FAR, distance to the nearest subway station ∩ FAR, transportation POI density ∩ FAR, and distance to the central business district (CBD) ∩ FAR made a higher contribution to ride-hailing ridership than the single-factor effect of FAR, which had the highest contribution compared with the other explanatory variables. The proposed grid scale can provide the basis for the partitioning management and scheduling optimization of ride-hailing. In the process of adjusting the ride-hailing demand, the ranking results of the importance and interaction of the built-environment explanatory variables offer valuable references for formulating the priority renewal order and proposing a scientific combination scheme of the built-environment factors.

Machine learning-based risk stratification for mortality in patients with severe aortic regurgitation.

The current guidelines recommend aortic valve intervention in patients with severe aortic regurgitation (AR) with the onset of symptoms, left ventricular enlargement, or systolic dysfunction. Recent studies have suggested that we might be missing the window of early intervention in a significant number of patients by following the guidelines. The overarching goal was to determine if machine learning (ML)-based algorithms could be trained to identify patients at risk for death from AR independent of aortic valve replacement (AVR). Models were trained with five-fold cross-validation on a dataset of 1035 patients, and performance was reported on an independent dataset of 207 patients. Optimal predictive performance was observed with a conditional random survival forest model. A subset of 19/41 variables was selected for inclusion in the final model. Variable selection was performed with 10-fold cross-validation using random survival forest model. The top variables included were age, body surface area, body mass index, diastolic blood pressure, New York Heart Association class, AVR, comorbidities, ejection fraction, end-diastolic volume, and end-systolic dimension, and the relative variable importance averaged across five splits of cross-validation in each repeat were evaluated. The concordance index for predicting survival of the best-performing model was 0.84 at 1 year, 0.86 at 2 years, and 0.87 overall, respectively. Using common echocardiographic parameters and patient characteristics, we successfully trained multiple ML models to predict survival in patients with severe AR. This technique could be applied to identify high-risk patients who would benefit from early intervention, thereby improving patient outcomes.

Yield and Properties Prediction Based on the Multicondition LSTM Model for the Solvent Deasphalting Process.

Solvent deasphalting (SDA) is a complex multiscale continuous process. The operation mode of the SDA process is not considered in the related data-driven model. Therefore, this paper proposes a time lag process prediction model with multiple operation modes to solve the above problem. First, based on random forests, the relative importance of initial input variables in the SDA process on DAO yield and Conradson carbon residual are studied and features are selected according to the results. Then, the stack denoising autoencoder (SDAE) is used to reconstruct the data and obtain the nonlinear mapping information of hidden layers of SDAE and achieve feature dimension reduction. SDAE can improve clustering accuracy of fuzzy c-means, and the operation mode of SDA process is accurately divided. Long short-term memory (LSTM) is used to establish a multicondition LSTM model. Compared with the traditional LSTM model, the multicondition LSTM model has a higher prediction accuracy with R 2 > 0.95. The sensitivity analyses of the properties of feed and operating conditions on DAO yield are consistent with the principle of two-phase countercurrent extraction in the SDA process. In addition, the benchmark test of the Tennessee Eastman process shows that the proposed method is also effective in the fault detection of other processes. Because the multicondition LSTM can predict the future process measurement data according to operating mode, it can better avoid the false alarm problem and predict the fault earlier.