Alternative Modeling Approach Research Articles

Enhancing speech emotion recognition with deep learning using multi-feature stacking and data augmentation

This study evaluates the effectiveness of data augmentation on 1D convolutional neural network (CNN) and transformer models for speech emotion recognition (SER) on the Ryerson audio-visual database of emotional speech and song (RAVDESS) dataset. The results show that data augmentation has a positive impact on improving emotion classification accuracy. Techniques such as noising, pitching, stretching, shifting, and speeding are applied to increase data variation and overcome class imbalance. The 1D CNN model with data augmentation achieved 94.5% accuracy, while the transformer model with data augmentation performed even better at 97.5%. This research is expected to contribute better insights for the development of accurate emotion recognition methods by using data augmentation with these models to improve classification accuracy on the RAVDESS dataset. Further research can explore larger and more diverse datasets and alternative model approaches.

Evaluation of a Simplified Modeling Approach for SEE Cross-Section Prediction: A Case Study of SEU on 6T SRAM Cells

Electrical models play a crucial role in assessing the radiation sensitivity of devices. However, since they are usually not provided for end users, it is essential to have alternative modeling approaches to optimize circuit design before irradiation tests, and to support the understanding of post-irradiation data. This work proposes a novel simplified methodology to evaluate the single-event effects (SEEs) cross-section. To validate the proposed approach, we consider the 6T SRAM cell a case study in four technological nodes. The modeling considers layout features and the doping profile, presenting ways to estimate unknown parameters. The accuracy and limitations are determined by comparing our simulations with actual experimental data. The results demonstrated a strong correlation with irradiation data, without requiring any fitting of the simulation results or access to process design kit (PDK) data. This proves that our approach is a reliable method for calculating the single-event upset (SEU) cross-section for heavy-ion irradiation.

Exploring potential mechanisms linking urban landscapes to instream stressors and macroinvertebrates in the Southeast Piedmont: Implications for watershed management

Urbanization has greatly degraded stream health, particularly in the rapidly urbanizing southeastern United States. Recent studies identified primary instream stressors affecting biological communities along urban gradients in the Southeast Piedmont. However, a comprehensive understanding of the complex mechanisms is still lacking, which impedes effective management actions. To improve our understanding, we investigated the potential pathways by which urbanization, as represented by the landscape development intensity index (LDII), and important environmental characteristics affected instream stressors and macroinvertebrate communities using structural equation modeling (SEM) and the U.S. Geological Survey Southeast Stream Quality Assessment dataset. We investigated the direct, indirect, and total effects of LDII and environmental characteristics on instream stressors and macroinvertebrate communities as well as the interactions of instream stressors. Moreover, we explored alternative reasonable relationships between variables by using an alternative model approach. We found that LDII had moderate to strong total effects on all instream stressors and macroinvertebrates with complex impact pathways revealed in SEM models, demonstrating the dominant impact of urban landscape development on stream ecosystems in the Southeast Piedmont. Flow disturbance and pesticides were major mediators transmitting most of LDII’s effects on macroinvertebrates. Furthermore, we found potential interactions among instream stressors, which were not widely discussed in previous studies. Compared with agricultural watersheds, urban watersheds represented different land use impact mechanisms. Our findings contributed to an improved understanding of urbanization’s impact on stream ecosystems and demonstrated the usefulness of an alternative model approach that provided insights into the underlying mechanisms. In addition, we demonstrated the utility of LDII in characterizing landscape development and improving SEM models. Finally, we explored the potential of SEM in global watershed management activities by discussing the practical implications of impact pathways and the innovative use of SEM as a decision-making tool.

Predictive Analysis of COVID-19 Epidemic in Thailand: Evaluating Control Lockdown Measures using LSTM Networks

This study addresses the critical objective of evaluating the effectiveness of non-pharmaceutical lockdown measures implemented during COVID-19 outbreaks in Thailand. Assessing the outcome of these measures provides valuable insight that can inform and guide response to future outbreaks. Utilizing a closed-loop forecasting model built on Long Short-Term Memory (LSTM) networks, the research focuses on achieving precise daily forecasts of COVID-19 cases. The methodology involves optimizing hyperparameters through grid-search and incorporating training data from other countries that implemented similar measures. The LSTM, configured with an optimal number of hidden processing units, utilizes past lagged data of daily infected cases as predictors to generate multi-step-ahead predicted values, which are subsequently used as predictors in a recursive approach. As a result, the predicted cases closely align with measured data, facilitating the estimation of the effective reproduction number (Reff) to assess the performance of lockdown measures. The effectiveness of the lockdown measures is quantified at different time intervals: 51%, 41%, and 23% one day after implementation, increasing to 84%, 98%, and 34% after one week, and reaching 96%, 99%, and 73% at the endpoint of the first, second, and fourth waves of infection, respectively. Throughout these waves, the final Reff remains below 1, indicating ongoing but controllable COVID-19, demonstrating the efficacy of the implemented lockdown measures. It is noted that these results are based on specific LSTM model, as the effectiveness of lockdown measures may vary with alternative modeling approaches. Therefore, the findings should be interpreted in the context of this LSTM-framework.

Enhancing physically-based hydrological modeling with an ensemble of machine-learning reservoir operation modules under heavy human regulation using easily accessible data

Dams and reservoirs have significantly altered river flow dynamics worldwide. Accurately representing reservoir operations in hydrological models is crucial yet challenging. Detailed reservoir operation data is often inaccessible, leading to relying on simplified reservoir operation modules in most hydrological models. To improve the capability of hydrological models to capture flow variability influenced by reservoirs, this study proposes a hybrid hydrological modeling framework, which combines a process-based hydrological model with a machine-learning-based reservoir operation module designed to simulate runoff under reservoir operations. The reservoir operation module employs an ensemble of three machine learning models: random forest, support vector machine, and AutoGluon. These models predict reservoir outflows using precipitation and temperature data as inputs. The Soil and Water Assessment Tool (SWAT) then integrates these outflow predictions to simulate runoff. To evaluate the performance of this hybrid approach, the Xijiang Basin within the Pearl River Basin, China, is used as a case study. The results highlight the superiority of the SWAT model coupled with machine learning-based reservoir operation models compared to alternative modeling approaches. This hybrid model effectively captures peak flows and dry period runoff. The Nash-Sutcliffe Efficiency (NSE) in daily runoff simulations shows substantial improvement, ranging from 0.141 to 0.780, with corresponding enhancements in the coefficient of determination (R2) by 0.098–0.397 when compared to the original reservoir operation modules in SWAT. In comparison to parameterization techniques lacking a dedicated reservoir module, NSE enhancements range from 0.068 to 0.537, and R2 improvements range from 0.027 to 0.139. The proposed hybrid modeling approach effectively characterizes the impact of reservoir operations on river flow dynamics, leading to enhanced accuracy in runoff simulation. These findings offer valuable insights for hydrological forecasting and water resources management in regions influenced by reservoir operations.

Integration of UAV Digital Surface Model and HEC-HMS Hydrological Model System in iRIC Hydrological Simulation—A Case Study of Wu River

This research investigates flood susceptibility in the mid- and downstream areas of Taiwan’s Wu River, historically prone to flooding in central Taiwan. The study integrates the Hydrologic Engineering Center—Hydrologic Modeling System (HEC-HMS) for flow simulations with unmanned aerial vehicle (UAV)-derived digital surface models (DSMs) at varying resolutions. Flood simulations, executed through the International River Interface Cooperative (iRIC), assess flood depths using diverse DSM resolutions. Notably, HEC-HMS simulations exhibit commendable Nash–Sutcliffe efficiency (NSE) exceeding 0.88 and a peak flow percentage error (PEPF) below 5%, indicating excellent suitability. In iRIC flood simulations, optimal results emerge with a 2 m resolution UAV-DSM. Furthermore, the study incorporates rainfall data at different recurrence intervals in iRIC flood simulations, presenting an alternative flood modeling approach. This research underscores the efficacy of integrating UAV-DSM into iRIC flood simulations, enabling precise flood depth assessment and risk analysis for flood control management.

Effectiveness of physical soil and water conservation measures under simulated and observed dataset in Ethiopia: Meta‐analysis

AbstractTo combat land degradation through conserving soil and water resources, Ethiopia has undergone vigorous implementation of soil and water conservation (SWC) measures, predominantly physical structures. To evaluate the effectiveness of these implemented structures, various researches have been conducted through using both measured and simulated approaches. This study was initiated to analyze the reliability of SWAT‐based simulated studies results on the effectiveness of SWC measures toward erosion reduction potential against the measured dataset using unpaired t‐test statistical analysis. In this study, 123 published papers were downloaded, of which 70 were found to be more related to the objective of the review. After applying four refining criteria, only 43 were found more compatible with the study objective and used for data analysis. Studies conducted by different scholars show up to 98% and 93.50% soil loss reduction effectiveness by SWC measures in measured and simulated approaches, respectively. The simulated SWC measure efficiency output was also evaluated with respect to the ground‐measured data of the same structure within the same agroecology class. Finally, the unpaired t‐test results indicated that the grass strip in the mid‐highland agroecology showed a significant difference from the measured one at a 95% significance level. However, none of the other measures showed significant differences between the measured and simulated datasets. Hence, the SWAT model can effectively simulate the effectiveness of physical SWC measures on soil loss reduction if it is well‐calibrated and validated with sediment yield data. Refining model parameters that can be sufficiently captured and represent biological measures (grass strip), incorporating additional field data for calibration and validation coupled with exploring alternative modeling approaches that efficiently simulate biological SWC was set as a recommendation to obtain an effective biological measure simulation.

Efficient modelling of surface roughness effects in additively manufactured materials

3D printing defects are often used to explain the mismatch between numerical predictions and experimental results of additively manufactured acoustic materials. Although they are often perceived as desirable, as in most cases, they improve the absorptive behaviour of the system, occurring gains are unplanned and uncontrollable. Among different 3D printing inaccuracies, surface roughness provides additional acoustic losses, mostly due to the enlargement of the boundary layer thickness. On the other hand, roughness is responsible for the shift of the operating frequency compared to the numerical predictions. The goal of this research is to provide efficient modelling methods to capture the effects of surface roughness arising from the additive manufacturing process. First, a direct modelling method will be discussed, which can be computationally expensive as it relies on an accurate geometrical representation of surface roughness. Second, an alternative indirect modelling approach will be presented, which mimics the presence of roughness by adjusting specific air parameters in the locations of interest of the smooth geometry to reduce the computational cost. The two strategies will be compared with each other and to the results of an experiment where additional roughness is added to target locations within the design.

Designing a Human-centered AI Tool for Proactive Incident Detection Using Crowdsourced Data Sources to Support Emergency Response

Time of incident reporting is a critical aspect of emergency response. However, the conventional approaches to receiving incident reports have time delays. Non-traditional sources such as crowdsourced data present an opportunity to detect incidents proactively. However, detecting incidents from such data streams is challenging due to inherent noise and data uncertainty. Naively maximizing detection accuracy can compromise spatial-temporal localization of inferred incidents, hindering response efforts. This article presents a novel human-centered AI tool to address the above challenges. We demonstrate how crowdsourced data can aid incident detection while acknowledging associated challenges. We use an existing CROME framework to facilitate training and selection of best incident detection models, based on parameters suited for deployment. The human-centered AI tool provides a visual interface for exploring various measures to analyze the models for the practitioner’s needs, which could help the practitioners select the best model for their situation. Moreover, in this study, we illustrate the tool usage by comparing different models for incident detection. The experiments demonstrate that the CNN-based incident detection method can detect incidents significantly better than various alternative modeling approaches. In summary, this research demonstrates a promising application of human-centered AI tools for incident detection to support emergency response agencies.

Cost-effectiveness of nivolumab versus surveillance for the adjuvant treatment of patients with urothelial carcinoma who are at high risk of recurrence: a US payer perspective

ABSTRACT Aim To evaluate the cost-effectiveness of adjuvant nivolumab compared with surveillance for the treatment of patients with high-risk muscle-invasive urothelial carcinoma (MIUC) after radical resection from a US healthcare payer perspective and to investigate the impact of alternative modeling approaches on the cost-effectiveness results. Material and Methods A four-state, semi-Markov model consisting of disease free, local recurrence, distant recurrence, and death health states was developed to investigate the cost-effectiveness of nivolumab compared with surveillance over a 30-year time horizon. The model used data from the randomized CheckMate 274 study (NCT02632409) and published literature to inform transitions among health states, and inputs on cost, utility, adverse event, and disease management. Scenario analyses were conducted to investigate the impact of model structure and key assumptions on the results. One-way deterministic and probabilistic sensitivity analysis were conducted to investigate the robustness of the results. Results Total expected costs were higher with nivolumab ($162,278) compared with surveillance ($63,027). Nivolumab was associated with improved survival (1.61 life-years gained compared with surveillance) and an incremental gain of 0.98 quality-adjusted life-years (QALYs). Although total treatment costs were higher for nivolumab, cost offsets were observed because of delayed or avoided recurrences and deaths experienced with nivolumab compared with observation. The incremental cost-effectiveness and cost-utility ratios were $61,462/life-year and $100,930/QALY. Limitations At the time of analysis, CheckMate 274 had limited follow-up on disease-free survival and no overall survival data. The limited evidence necessitated assumptions on modeling survival after each type of recurrence. Conclusions Nivolumab is estimated to be a life-extending and cost-effective option for adjuvant treatment of MIUC for patients who are at high risk of recurrence after undergoing radical resection in the United States. Using a threshold of $150,000/QALY, the cost-effectiveness conclusions remained consistent across the scenario and sensitivity analyses conducted.

Matching decision support modeling frameworks to disease emergence stages and associated management objectives

AbstractWildlife disease management decisions often require rapid responses to situations that are fraught with uncertainty. By recognizing that management is implemented to achieve specific objectives, resource managers and science partners can identify an analysis technique and develop a monitoring plan to evaluate management effectiveness. For emerging infectious diseases, objectives may take several distinct forms, dependent on the perceived stage of disease emergence (i.e., pre‐epidemic, early outbreak, mid‐epidemic, and endemic), the expected rate of spread, and the anticipated effect of the disease on host populations. Identifying modeling techniques and metrics that are linked to management objectives will require early and consistent communication between managers and science partners. We link modeling approaches that can be used to forecast and evaluate the performance of intervention strategies with a range of disease management objectives. Our aim is to help scientists recognize alternative modeling approaches which may better align with different forms of disease management objectives, and to help managers evaluate the relevance of proposed modeling approaches to their specified objectives for disease management. Recognizing that disease management objectives can take different forms, and thus require different modeling approaches, can help wildlife disease response teams (i.e., natural resource managers, scientists, and stakeholders working collaboratively) better prepare and respond to disease threats.

Forecasting disease trajectories in critical illness: comparison of probabilistic dynamic systems to static models to predict patient status in the intensive care unit

ObjectiveConventional prediction models fail to integrate the constantly evolving nature of critical illness. Alternative modelling approaches to study dynamic changes in critical illness progression are needed. We compare static risk...

On understanding mountainous carbonate basins of the Mediterranean using parsimonious modeling solutions

Abstract. The study aims to demonstrate that an effective solution can be implemented for modeling complex carbonate basins, in the situation of limited data availability. Considering the alternative modeling approaches under circumstances of data shortage is more significant knowing the vulnerability and effectiveness of these kinds of basins to drought and climate change conditions. In this regard, a hybrid approach that combines time series analysis and reservoir modeling is proposed to describe behavior in carbonate basins. Time series analysis estimates the contributing area and response time of the fractured carbonate system beyond the catchment's hydrographic boundaries. The results obtained align with previous literature-based field surveys. This information is then used to develop a conceptual reservoir system using the GEOframe modeling system. The model is validated using in situ discharge observations and Earth observations (EO) data on evapotranspiration and snow. Model reliability is assessed using traditional goodness of fit indicators, hydrological signatures, and a novel statistical method based on empirical conditional probability. This approach enables detailed analysis and investigation of water budget components in Mediterranean carbonate catchments, highlighting their response to significant precipitation deficits. Overall, our results demonstrate that flows from carbonate rock areas outside the hydrographic boundaries significantly impact the water budget of the upper Nera River. The storage capacity of the carbonate basin plays a crucial role in sustaining river discharge during drought years. In a single dry year, meteorological drought is considerably attenuated, while in subsequent dry years, it is slightly intensified. Multi-year droughts result in slower recovery due to the time required for precipitation to replenish the depleted storage that supported river discharge in previous dry years. This unique behavior makes these basins particularly vulnerable to the more severe and frequent drought episodes expected under future climate change.

Quantifying pesticide emissions for drift deposition in comparative risk and impact assessment

Estimating emissions of chemical pesticides used in agriculture is an essential component in evaluating the potential toxicity-related impacts on humans and ecosystems in various comparative risk and impact assessment frameworks, such as life cycle assessment, environmental footprinting, absolute environmental sustainability assessment, chemical substitution, and risk prioritization. Emissions related to drift deposition—usually derived from drift experiments—can reach non-target areas, and vary as a function of crop characteristics and application technique. We derive cumulative drift deposition fractions for a wide range of experimental drift functions for use in comparative and mass-balanced approaches. We clarify that cumulative drift deposition fractions require to integrate the underlying drift functions over the relevant deposition area and to correct for the ratio of deposition area to treated field area to arrive at overall mass deposited per unit mass of applied pesticide. Our results show that for most crops, drift deposition fractions from pesticide application are below 0.03 (i.e. 3% of applied mass), except for grapes and fruit trees, where drift fractions can reach 5% when using canon or air blast sprayers. Notably, aerial applications on soybeans can result in significantly higher drift deposition fractions, ranging from 20% to 60%. Additionally, varying the nozzle position can lead to a factor of five differences in pesticide deposition, and establishing buffer zones can effectively reduce drift deposition. To address remaining limitations in deriving cumulative drift deposition fractions, we discuss possible alternative modelling approaches. Our proposed approach can be implemented in different quantitative and comparative assessment frameworks that require emission estimates of agricultural pesticides, in support of reducing chemical pollution and related impacts on human health and the environment.

Generative spatial generalized dissimilarity mixed modelling (spGDMM): An enhanced approach to modelling beta diversity

Abstract Turnover, or change in the composition of species over space and time, is one of the primary ways to define beta diversity. Inferring what factors impact beta diversity is not only important for understanding biodiversity processes but also for conservation planning. At present, a popular approach to understanding the drivers of compositional turnover is through generalized dissimilarity modelling (GDM). We argue that the current GDM approach suffers several limitations and provide an alternative modelling approach that remedies these issues. We propose using generative spatial random effects models implemented in a Bayesian framework. We offer hierarchical specifications to yield full regression and spatial predictive inference, both with associated full uncertainties. The approach is illustrated by examining dissimilarity in three datasets: tree survey data from Panama's Barro Colorado Island (BCI), plant occurrence data from southwest Australia and plant abundance surveys from the Greater Cape Floristic Region (GCFR) of South Africa. We select a best model using out‐of‐sample predictive performance. We find that the form of the best model differs across the three datasets, but our models provide performance ranging from comparable to significant improvement over GDMs. Within the GCFR, the spatial random effects play a more important role in the modelling than all the environmental variables. We have proposed a model that provides several improvements to the current GDM framework. This includes advantages such as a flexible spatially varying mean function, spatial random effects that capture dependence unaccounted for by explanatory variables, and spatially heterogeneous variance structure. All these features are offered in a model that can adequately handle a large incidence of total dissimilarity through ‘one‐inflation’, as would be expected from highly biodiverse areas with steep turnover gradients.

Digital rock advances from a material point method approach for simulation of frame moduli and a sedimentary petrology-inspired method for creation of synthetic samples through simulation of deposition and diagenesis

We compare hydromechanical simulation results that use two alternative sources of 3D digital rock input: micro-CT analysis and “synthetic rocks” created by using a newly developed process simulation methodology that more rigorously reflects knowledge from sedimentary petrology compared with previous efforts. We evaluate the performance of these alternative representations using St. Peter Sandstone samples where “dry” static bulk modulus ( K) and shear modulus ( G) are simulated using a new extension of the material point method that resolves contacts using high-resolution surface meshes and considers three alternative contact modeling approaches: “purely frictional,” “fully bonded,” and “cohesive zones.” We evaluate the model performance on two samples from the data set with multiple static moduli measurements (sample 1_2: porosity 24.6 vol%, K 10.2–14.7 GPa, and G 11.6–14.0 GPa; sample 11_2: porosity 12.4 vol%, K 13.5–24.6 GPa, and G 12.8–17.9 GPa). Purely frictional results underpredict measured modulus values, whereas fully bonded results overpredict them. Measured values are most closely approximated by results with cohesive zones that consider sets of discrete spring-like features at contacts. In contrast, shear modulus results from finite-element model simulations on structured grids tend to be significantly greater than measured values, particularly for samples with <18 vol% porosity. Permeability values from digital rock-physics simulations for the studied samples are within factors of 2–5 of conventional core analysis measurements (2860 and 58 mD for samples 1_2 and 11_2, respectively). We determine that the process modeling approach (1) accurately reproduces the measured rock microstructure parameters from thin-section analysis, (2) leads to simulation results for dry static moduli and permeability with accuracy comparable to simulations that use micro-CT samples, and (3) provides a rigorous basis for predicting diagenetically induced variations in hydromechanical properties over the range from unconsolidated sand to indurated rock.

Clarifying status of DNNs as models of human vision.

On several key issues we agree with the commentators. Perhaps most importantly, everyone seems to agree that psychology has an important role to play in building better models of human vision, and (most) everyone agrees (including us) that deep neural networks (DNNs) will play an important role in modelling human vision going forward. But there are also disagreements about what models are for, how DNN-human correspondences should be evaluated, the value of alternative modelling approaches, and impact of marketing hype in the literature. In our view, these latter issues are contributing to many unjustified claims regarding DNN-human correspondences in vision and other domains of cognition. We explore all these issues in this response.

Abstract 12291: Impact of History of Mental Health Disorder on In-Hospital Mortality in Patients With Heart Failure. A Population-Based Cohort Study

Introduction: Mental health disorders are associated with lower life expectancy and patients with mental disorders are at an increased risk of heart failure. Data on the prognostic impact of mental disorders on heart failure outcomes is limited and inconclusive. Hypothesis: Mental health disorders are associated with increased risk of in-hospital mortality in patients with heart failure. Methods: The study population were patients aged ≥18 years in acute care hospitals with heart failure as principal diagnosis in Texas for years 2016 through 2021. Three prespecified models were applied to estimate the association between mental disorders and in-hospital mortality. Multilevel, multivariable logistic regression with propensity adjustment was the primary analysis approach with propensity score matching and multivariable logistic regression without propensity adjustment applied as alternative analyses. Additional analyses were completed for patient subgroups, categories of mental disorders, and classifications of heart failures. Results: Out of 513,565 heart failure hospitalizations, 103,675 (20.2%) had mental disorders. Heart failure hospitalizations with vs without mental disorders were more commonly female (61.3% vs 46.0%), white (55.3% vs 46.9%), and obese (31.6% vs 26.4%), but had less frequent need for mechanical ventilation (1.8% vs 2.1%); p < 0.0001 for each. Unadjusted in-hospital mortality among heart failure hospitalizations with and without mental disorders was 1.7% vs 2.2%. On adjusted analyses, mental disorders remained associated with lower odds of in-hospital mortality (adjusted odds ratio 0.8650 [95% CI 0.8149-0.9183]). The results were similar for the alternative modeling approaches, subgroup analyses, and classifications of heart failure. Conclusion: Mental disorders were associated with substantially lower risk of in-hospital mortality for patients with a principal diagnosis of heart failure.

Abstract 14620: Association of Rural Residence With In-Hospital Mortality for Heart Failure Hospitalizations in Texas: A Population-Based Cohort Study

Background: Rural residents with heart failure may be at higher risk of mortality due to disparities in access to health care services. Patterns of urban-rural disparity in health care resources may vary across the United States and population level analyses of the impact of rural residence on hospitalized patients with heart failure in Texas are not conclusive. Hypothesis: Rural residence is associated with increased risk of in-hospital mortality in patients with heart failure. Methods: We used publicly available, statewide, and deidentified hospitalization records to study patients aged ≥18 years with heart failure as principal diagnosis in Texas for years 2016 through 2021. Three models estimated the association between rural residence and in-hospital mortality. Multilevel logistic regression with propensity adjustment was the primary analysis approach with propensity score matching and multivariable logistic regression without propensity score matching applied as alternative analyses. Additional analyses were completed for patient subgroups and for categories of heart failures. Results: Out of 460,798 heart failure hospitalizations (HFH), 61,480 (13.3%) were classified as rural residents. HFH with vs without rural residence were older (70.4% vs 64.8% aged ≥ 65years), more often white (63.5% vs 46.9%), less frequently had private insurance (26.2% vs 35.4%) and were less often admitted to teaching hospitals (19.5% vs 28.9%); p < 0.0001 for each. Unadjusted in-hospital mortality among HFH with and without rural residence was 2.5% vs 2.1%. On adjusted analyses, rural residence remained associated with higher odds of in-hospital mortality (adjusted odds ratio 1.1685 (95% CI [1.0878 - 1.2554]). The results were similar for the alternative modeling approaches, subgroup analyses, and categories of heart failures. Conclusions: Rural residence in Texas was associated with higher risk of in-hospital mortality in patients with heart failure.

Predictive Modeling for Highway Pavement Rutting: A Comparative Analysis of Auto-Machine Learning and Structural Equation Models

Highway pavements deteriorate over time as successive wheel loads cause rutting, cracking, texture loss, and so forth. Design standards and pavement performance models account for some of the known contributory factors, such as levels of traffic and vehicle composition. However, such models are limited in their predictive power, and highway authorities must conduct regular pavement condition surveys rather than relying on the standard deterioration models alone. The ways in which multiple factors affect pavement deterioration, including rutting, are complex and are believed to include feedback loops where rutting then influences driving position, exacerbating the rutting levels. Standard regression models are not well suited to representing such complex causal mechanisms. This paper compares two alternative modeling approaches, structural equation models and auto-machine learning, and evaluates the predictive ability and practicalities of each. The findings indicate that auto-machine learning (AutoML) may be superior in its predictive ability. However, the “black box” nature of AutoML results makes them potentially less useful to practitioners. A process of using machine learning to help inform a structural equation model is proposed.