Model For Predicting Performance Research Articles

A Predictive Model for Academic Performance of Students in MAT101 at Federal University Birnin Kebbi

The research addresses the problem of low academic performance of students taking General Mathematics 1 (MAT101) at the Federal University Birnin Kebbi in Nigeria. MAT101 is a foundational course for several science-related disciplines. However, many students fail to understand its concepts, translating to one of the lowest academic performances, leading to high discontinuation of students’ studies in most science and related fields. To this end, the problem is addressed through a predictive model using data science and machine learning techniques to identify students who are academically at risk. It seeks to answer the appropriateness of the selected statistical and data mining approaches, whether or not they are predictor variables, and validating the model developed. The significance of this study lies in its potential to help instructors identify at-risk students and improve overall performance in MAT101. The research uses data collected from academic records and conducts exploratory data analysis to determine the factors influencing students’ performance. Several classifiers such as Decision Tree and Random Forest were analyzed by evaluating their accuracy. In detail, the difference in the accuracy for the Decision Tree algorithm was 52% and for the Random Forest was 67%. Based on achieved indicators, the Random forest algorithm is more effective in this type of predictive modeling. However, the choice of algorithm depends on specific task requirements, considering factors like speed and accuracy.

Validation of predictive performance models for supersonic gas-jet nozzles at the Laboratory for Laser Energetics.

We present results characterizing the neutral-density distributions produced by the supersonic nozzles used in experiments on the OMEGA-60 and OMEGA-EP laser systems at the University of Rochester's Laboratory for Laser Energetics (LLE). Axisymmetric Fluent® simulations using LLE nozzle specifications capture the viscous effects, gas expansion, and shock waves that complicate flow predictions for offsets above the nozzle exit. These simulations show good agreement with neutral-density measurements obtained using a four-wave shearing interferometer. An analytical form is given for the plateau length. Fits to simulation data for boundary layer thickness, mean plateau density, and density ramps are given as functions of nozzle offset and nozzle backing pressure for a number of nozzles and gases.

Thermal management of ultra high concentrator photovoltaic cells: Analysing the impact of sintered porous media microchannel heat sinks

Implementing advanced cooling strategies is essential to maintaining recommended operation conditions, enhancing energy efficiency, and extending the reliability of ultra-high concentrator photovoltaics. This study introduces a new application of sintered porous media as an effective and superior thermal management solution for ultra-high concentrator photovoltaic systems. Furthermore, predictive models for thermal and electrical performance are developed.A comprehensive three-dimensional computational fluid dynamics model was developed and verified against experimental work reported to investigate the performance of a novel sintered porous media for thermal management in ultra-high concentrator photovoltaics. Furthermore, the response surface methodology coupled with analysis of variance was used to analyze and comprehensively optimize the thermal management system. The main objective is to develop a predictive model for such systems equipped with a porous cooling system that predicts performance with a concentration ratio of up to 20000 Suns. Predictive models were formulated to outline the relationships between various independent variables: Reynolds number, concentration ratio, direct normal irradiance, wind speed, optical efficiency, fluid inlet temperature, and dependent performance metrics.This study highlights the effectiveness of sintered porous media in heat dissipation and efficiency enhancement in ultra-high concentrator photovoltaics thermal management. The results demonstrate that using sintered porous media compared to conventional channel heatsinks significantly enhances cell thermal management. Specifically, this approach reduces cell temperature by 15.3%, 8.97%, and 2% at concentration ratios of 20000, 10000, and 2000, respectively. Concurrently, implementing the new sintered porous media heatsink improves cell efficiency by 3.1%, 1.5%, and 0.3% for the same respective concentration ratio values.

Investigation of Corrosion Inhibition Efficiency of Pyridine-Quinoline Compounds through Machine Learning

Corrosion in materials is a significant concern for the industrial and academic fields because corrosion causes enormous losses in various fields such as the economy, environment, society, industry, security, safety, and others. Currently, material damage control using organic compounds has become a popular field of study. Pyridine and quinoline stand out as corrosion inhibitors among a myriad of organic compounds because they are non-toxic, inexpensive, and effective in a variety of corrosive environments. Experimental investigations in developing various candidate potential inhibitor compounds are time and resource-intensive. In this work, we use a quantitative structure-property relationship (QSPR)-based machine learning (ML) approach to investigate support vector machine (SVR), random forest (RF), and k-nearest neighbors (KNN) algorithms as predictive models of inhibition performance. (Inhibition efficiency) corrosion of pyridine-quinoline derivative compounds as corrosion inhibitors on iron. We found that the RF model showed the best predictive ability based on the coefficient of determination (R2) and root mean squared error (RMSE) metrics. Overall, our study provides new insights regarding the ML model in predicting corrosion inhibition on iron surfaces.

Ensemble Learning Approach for Developing Performance Models of Flexible Pavement

This research utilizes the Long-Term Pavement Performance database, focusing on devel-oping a predictive model for flexible pavement performance in the Southern United States. Analyzing 367 pavement sections, this study investigates crucial factors influencing asphaltic concrete (AC) pavement deterioration, such as structural and material components, air voids, compaction density, temperature at laydown, traffic load, precipitation, and freeze–thaw cycles. The objective of this study is to develop a predictive machine learning model for AC pavement wheel path cracking (WpCrAr) and the age at which cracking initiates (WpCrAr) as performance indicators. This study thoroughly investigated three ensemble machine learning models, including random forest, extremely randomized trees (ETR), and extreme gradient boosting (XGBoost). It was observed that XGBoost, optimized using Bayesian methods, emerged as the most effective among the evaluated models, demonstrating good predictive accuracy, with an R2 of 0.79 for WpCrAr and 0.92 for AgeCrack and mean absolute errors of 1.07 and 0.74, respectively. The most important features influencing crack initiation and progression were identified, including equivalent single axle load (ESAL), pavement age, number of layers, precipitation, and freeze–thaw cycles. This paper also showed the impact of pavement material combinations for base and subgrade layers on the delay of crack initiation.

Ultra-High-Performance Alkali-Activated Concrete: Effect of Waste Crumb Rubber Aggregate Proportions on Tensile and Flexural Properties

The declining availability of natural sand resources and the significant carbon footprint associated with the extensive use of cement are posing severe limitations on the advancement and application of ultra-high-performance concrete (UHPC). In this study, waste tyre-derived recycled crumb rubber particles (CR) were employed to replace quartz sand, and an alkali-activated cementitious material was used to produce waste tyre-alkali-activated UHPC (T-UHPAC). The influence of different CR replacement ratios (0%, 5%, 20%, 35%, 50%) on the tensile and flexural performance of T-UHPAC was investigated, and a predictive model for the stress–strain response considering the CR replacement ratio was established. An optimization method for improving the tensile and flexural performance of T-UHPAC was proposed. The results indicate that the effect of rough-surfaced CR on the interfacial properties of concrete differs from that of smooth quartz sand. A CR replacement ratio exceeding 35% led to a reduction in both the tensile and flexural strengths of UHPAC, while a replacement ratio at or below 20% resulted in a superior tensile and flexural performance of T-UHPAC. The established predictive model for tensile performance accurately forecasts the stress–strain behaviour of T-UHPAC under varying CR replacement ratios, with the accuracy improving as the CR replacement ratio increases. By utilizing CR to replace quartz sand in proportions not exceeding 20%, the production of low-carbon UHPC with exceptional comprehensive mechanical properties is achievable. Moreover, the development of T-UHPAC through the comprehensive utilization of waste tyres presents a promising and innovative approach for the low-carbon and cost-effective production of UHPC, thereby facilitating the sustainable development of natural resources. This research represents a significant step towards the widespread adoption and application of UHPC and thus holds substantial importance.

Uncertainty Quantification of Data-driven Quality Prediction Model For Realizing the Active Sampling Inspection of Mechanical Properties in Steel Production

Pre-production quality defect inspection is a crucial step in industrial manufacturing, and many traditional inspection strategies suffer from inefficiency issues. This is especially true for tasks such as mechanical performance testing of steel products, which involve time-consuming processes like offline sampling, specimen preparation, and testing. The inspection volume significantly impacts the production cycle, inventory, yield, and labor costs. Constructing a data-driven model for predicting product quality and implementing proactive sampling inspection based on the prediction results is an appealing solution. However, the prediction uncertainty of data-driven models poses a challenging problem that needs to be addressed. This paper proposes an active quality inspection approach for steel products based on the uncertainty quantification in the predictive model for mechanical performance. The objective is to reduce both the sampling frequency and the omission rate on the production site. First, an ensemble model based on improved lower and upper bound estimation is established for interval prediction of mechanical performance. The uncertainty of the specific value prediction model is quantitatively estimated using interval probability distributions. Then, a predictive model for the mechanical performance failure probability is built based on the prediction interval size and probability distribution. By determining an appropriate probability threshold, the trade-off between prediction accuracy and defect detection accuracy (recall rate) is balanced, enabling the establishment of an active sampling strategy. Finally, this functionality is integrated into the manufacturing execution system of a steel factory, realizing a mechanical performance inspection approach based on proactive sampling. The proposed approach is validated using real production datasets. When the probability threshold is set to 30%, the prediction accuracy and recall rate for failure mechanical performance samples are 75% and 100%, respectively. Meanwhile, the sampling rate is only 5.33%, while controlling the risk of omission. This represents a 50% reduction in sampling rate compared to the inspection rules commonly used in actual production. The overall efficiency of product quality inspection is improved, and inspection costs are reduced.

Weighting the structural connectome: Exploring its impact on network properties and predicting cognitive performance in the human brain.

Brain function does not emerge from isolated activity, but rather from the interactions and exchanges between neural elements that form a network known as the connectome. The human connectome consists of structural and functional aspects. The structural connectome (SC) represents the anatomical connections, and the functional connectome represents the resulting dynamics that emerge from this arrangement of structures. As there are different ways of weighting these connections, it is important to consider how such different approaches impact study conclusions. Here, we propose that different weighted connectomes result in varied network properties, and while neither superior the other, selection might affect interpretation and conclusions in different study cases. We present three different weighting models, namely, number of streamlines (NOS), fractional anisotropy (FA), and axon diameter distribution (ADD), to demonstrate these differences. The later, is extracted using recently published AxSI method and is first compared to commonly used weighting methods. Moreover, we explore the functional relevance of each weighted SC, using the Human Connectome Project (HCP) database. By analyzing intelligence-related data, we develop a predictive model for cognitive performance based on graph properties and the National Institutes of Health (NIH) toolbox. Results demonstrate that the ADD SC, combined with a functional subnetwork model, outperforms other models in estimating cognitive performance.

Glacier: guided locally constrained counterfactual explanations for time series classification

In machine learning applications, there is a need to obtain predictive models of high performance and, most importantly, to allow end-users and practitioners to understand and act on their predictions. One way to obtain such understanding is via counterfactuals, that provide sample-based explanations in the form of recommendations on which features need to be modified from a test example so that the classification outcome of a given classifier changes from an undesired outcome to a desired one. This paper focuses on the domain of time series classification, more specifically, on defining counterfactual explanations for univariate time series. We propose Glacier, a model-agnostic method for generating locally-constrained counterfactual explanations for time series classification using gradient search either on the original space or on a latent space that is learned through an auto-encoder. An additional flexibility of our method is the inclusion of constraints on the counterfactual generation process that favour applying changes to particular time series points or segments while discouraging changing others. The main purpose of these constraints is to ensure more reliable counterfactuals, while increasing the efficiency of the counterfactual generation process. Two particular types of constraints are considered, i.e., example-specific constraints and global constraints. We conduct extensive experiments on 40 datasets from the UCR archive, comparing different instantiations of Glacier against three competitors. Our findings suggest that Glacier outperforms the three competitors in terms of two common metrics for counterfactuals, i.e., proximity and compactness. Moreover, Glacier obtains comparable counterfactual validity compared to the best of the three competitors. Finally, when comparing the unconstrained variant of Glacier to the constraint-based variants, we conclude that the inclusion of example-specific and global constraints yields a good performance while demonstrating the trade-off between the different metrics.

Combined advanced oxidation dye-wastewater treatment plant: design and development with data-driven predictive performance modeling

The recalcitrant nature of the industrial dyes poses a significant challenge to existing treatment technologies due to the stringent environmental regulations. This combined with the inefficiency of a single treatment method has led to the implementation of the combination of primary, secondary, and tertiary treatment processes, which fails during complex secondary aeration processes due to variable pH loads of industrial effluent wastewater. This article presents a modified design methodology of a pilot-scale micro-pre-treatment unit using a solar-triggered advanced oxidation process reactor that both effectively controls the influent variability at the source and mitigates textile effluents for making the discharge reusable for different industrial purposes. The proposed modified combination technique of controlled serial processes inclusive of primary, secondary, and tertiary treatment steps with ZnO/ZnO-GO NanoMat-based advanced oxidation process demonstrates complete remediation of industrial grade effluent with effective reuse of the discharge. Further, a reliable prediction model for estimating water quality parameter using machine learning models are proposed. Multi-linear regression and Artificial Neural network modeling provide simple, accurate, and robust prediction capabilities, which are evaluated for the efficiency of the processes. The generated prediction models capture the output parameters within an acceptable level of accuracy (Radj2>0.90)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$({{\\boldsymbol{R}}}_{{adj}}^{{\\bf{2}}}\\, >\\, 0.90)$$\\end{document} and allow compliance with the discharge Inland Water Discharge Standards (IWDS).

Using Support Vector Regression Kernel Models for Cricket Performance Prediction in the Womens Premier League 2024

Background. The interest in women’s premier league cricket has caused the need for advanced analytics to understand the multifaceted dynamics of the sport. Study Purpose. This study aimed to contribute to sports analytics by assessing the efficacy of Support Vector Regression (SVR) kernel models in predicting the most valuable player. Such research methods as ANOVA, Bessel function, and Inverse MultiQuadratic kernel application have been deliberately chosen for their diverse mathematical approaches, aligning with the nuanced intricacies of women's premier league cricket. Materials and methods. Player performance was analyzed by using the following study methods: ANOVA, Bessel function and Inverse MultiQuadratic kernel application. The data, sourced from espncricinfo.com and the International Cricket Council, includes essential metrics for five teams. Rigorous preprocessing techniques, such as imputation and outlier removal, enhance data reliability, ensuring robust predictive models. Results. The application of the Inverse MultiQuadratic kernel exhibits exceptional predictive performance, surpassing ANOVA and Bessel function models. The kernels radial basis function proves effective in capturing the intricate dynamics of women’s premier league cricket. The findings underscore the suitability of kernel method for predicting standout performers in the Womenʼs Premier League 2024 season. Conclusions. The study revealed the dynamic interplay between sports analytics and machine learning in women's premier league cricket. The application of the Inverse MultiQuadratic kernel stands out as the most effective model, providing key insights into player predictions. This emphasizes the continual integration of advanced analytical techniques to enhance our understanding of the evolving landscape of women’s premier league cricket. As the sport gains prominence on the global stage, such analytical endeavors become imperative for strategic decision-making and sustained growth.

Exploring school factors of academic achievement in Macau: The application of educational data mining in Progress in International Reading Literacy Study (PIRLS) 2016 assessment

International large-scale assessments, provide structured and static data, and due to their extensive databases, they can be considered as a valuable resource for Big Data in Education. In this paper, we propose an educational data mining approach to detect and analyze factors linked to academic performance in Macau schools using data from the Progress in International Reading Literacy Study (PIRLS) 2016. We conducted a secondary data analysis based on a set of socioeconomic, process, and outcome variables from PIRLS and other sources, and built decision trees to obtain a predictive model of school performance. By doing so, we were able to identify the school and student-level variables that are most significant in predicting student performance in Macau. These findings will be useful for informing educational policy decisions and shedding light on the causes of poor performance in Macau schools. Overall, our study highlights the potential of educational data mining approaches in analyzing large-scale assessment data and generating insights for educational research and practice.

Validity of submaximal aerobic capacity and strength tests in firefighters.

Typically, the fitness of UK firefighters is assessed via submaximal estimate methods due to the low demands on time, money, expertise and equipment. However, the firefighter-specific validity of such testing in relation to maximum aerobic capacity (V˙O2max) and particularly muscular strength is not well established. To examine the validity of submaximal methods to estimate V˙O2max and maximal strength in operational firefighters. Twenty-two full-time operational firefighters (3 female) completed same-day submaximal (Chester Step Test; CST) and maximal (treadmill) assessments of V˙O2max, with a sub-sample of 10 firefighters (1 female) also completing submaximal and maximal back-squat (i.e. one repetition maximum; 1RM) assessments. All participants then completed the Firefighter Simulation Test (FFST) within 2-4 days. CST underestimated actual V˙O2max by 1.4 ml·kg-1·min-1 (~3%), although V˙O2max values were positively correlated (r = 0.61, P < 0.01) and not significantly different. Estimated V˙O2max values negatively correlated with FFST performance (r = -0.42). Predicted 1RM underestimated actual 1RM by ~2%, although these values were significantly correlated (r = 0.99, P < 0.001) and did not significantly differ. The strongest predictive model of FFST performance included age, body mass index, and direct maximal measures of 1RM and V˙O2max. Submaximal back-squat testing offers good validity in estimating maximum firefighter strength without exposure to the fatigue associated with maximal methods. The CST provides a reasonably valid and cost-effective V˙O2max estimate which translates to firefighting task performance, although the error observed means it should be used cautiously when making operational decisions related to V˙O2max benchmarks.

Use of Human-Automation Taxonomies for System Modeling

Skraaning and Jamieson indicate the importance of developing a definition of automation failure in order to guide development of human performance models to predict human-automation interactions. They go on to develop a taxonomy including four categories: Elementary automation failures; Systemic automation failures; Human-automation breakdowns; and Human and organizational slips and misconceptions. They suggest that the latter category be excluded from classes of automation failures. They then proceed to identify important factors within each category. This paper presents a complementary taxonomy designed to identify factors contributing to automation-related failures in order to develop probes to elicit predictions of failure stories for some proposed new system design from a range of experts. While there is overlap between these two taxonomies, there are important differences. It is suggested that they could be integrated or used jointly both for knowledge elicitation and to identify requirements regarding needed capabilities of predictive performance models.

Predicting forward osmosis performance with synthesized polyamide-based membrane: An integrated machine learning (MATLAB and ANN) and economic analysis framework

Forward Osmosis (FO) has emerged as a highly promising membrane-based water treatment process owing to its intrinsic advantages, such as low energy requirements and the potential for high water recovery. The lab-prepared highly efficient thin film composite (TFC) FO membranes that were used in this research work. The resultant polyamide-based FO membrane showed exceptional performance, notably achieving a high-water flux of 60.94 ± 3.5 L/m2.h and a constrained solute flux of 1.52 ± 0.08 g/m2.h. Machine learning methodologies are strategically employed in this study to construct predictive models for FO membrane performance. Notably, the accuracy of the van't Hoff equation's linear assumption is significantly enhanced by introducing an osmotic pressure calculation based on water activity, considering the nuanced effects of external and internal concentration polarization. The research employs MATLAB software alongside an artificial neural network (ANN) to develop predictive models. MATLAB facilitates the creation of a comprehensive theoretical model, offering insights into forecasting water flux through FO membranes by systematically analyzing diverse membrane parameters and their impact on performance. Concurrently, the ANN model was employed to predict the permeate flux for the laboratory-scale FO membrane system, utilizing the input parameters within the network's training range (R2 > 0.91). An economic assessment is also conducted to estimate the projected membrane cost, a pivotal determinant of system viability. Collectively, these investigations yield valuable insights into the intricacies of FO membrane design and optimization, elucidating their performance dynamics in water treatment and allied applications. This research significantly contributes to the ongoing quest to develop more efficient and precise membrane processes.

The effect of covariates on Soil Organic Matter and pH variability: a digital soil mapping approach using random forest model

ABSTRACT This research focuses on understanding the spatial variation of Soil Organic Matter (SOM) and pH levels in the North of Morocco. The study employs a comprehensive approach to enhance predictive modelling, incorporating the Boruta algorithm for effective environmental covariates selection and optimizing model parameters through hyperparameter optimization. Utilizing a Random Forest (RF) model with remote sensing indices and topographic features, the research predicts SOM and pH to identify key contributors to their spatial variability. SOM prediction saw significant success, with a notable correlation to remote sensing indices such as the RVI, NDVI, and TNDVI. These indices, indicative of vegetation health and productivity, emerged as primary influencers of SOM. In comparison, the influence of topographic features like elevation, slope, and aspect was found to be less significant. Conversely, predicting pH was challenging due to the minimal spatial variability within the dataset. Addressing this limitation could involve dataset expansion or alternative models for low-correlated data handling. Despite the RF model’s limited efficacy in pH prediction, an observable correlation between SOM and pH was identified, consistent with prior research. Areas with higher SOM exhibited lower pH values, indicating relative soil acidification from organic matter decomposition. The study’s RF model demonstrated potential in SOM prediction using remote sensing indices, but enhancing pH prediction is essential. Future research may explore dataset expansion, diverse sampling, or testing alternative predictive models for better performance with low-correlated datasets. The study offers valuable insights for advanced predictive model development and enriches understanding of soil management practices.

MATEO: intermolecular α-amidoalkylation theoretical enantioselectivity optimization. Online tool for selection and design of chiral catalysts and products.

The enantioselective Brønsted acid-catalyzed α-amidoalkylation reaction is a useful procedure is for the production of new drugs and natural products. In this context, Chiral Phosphoric Acid (CPA) catalysts are versatile catalysts for this type of reactions. The selection and design of new CPA catalysts for different enantioselective reactions has a dual interest because new CPA catalysts (tools) and chiral drugs or materials (products) can be obtained. However, this process is difficult and time consuming if approached from an experimental trial and error perspective. In this work, an Heuristic Perturbation-Theory and Machine Learning (HPTML) algorithm was used to seek a predictive model for CPA catalysts performance in terms of enantioselectivity in α-amidoalkylation reactions with R2 = 0.96 overall for training and validation series. It involved a Monte Carlo sampling of > 100,000 pairs of query and reference reactions. In addition, the computational and experimental investigation of a new set of intermolecular α-amidoalkylation reactions using BINOL-derived N-triflylphosphoramides as CPA catalysts is reported as a case of study. The model was implemented in a web server called MATEO: InterMolecular Amidoalkylation Theoretical Enantioselectivity Optimization, available online at: https://cptmltool.rnasa-imedir.com/CPTMLTools-Web/mateo . This new user-friendly online computational tool would enable sustainable optimization of reaction conditions that could lead to the design of new CPA catalysts along with new organic synthesis products.

ZrO2-reinforced polymer-matrix composites used for thermal protection systems of ultra-high temperature aerospace propulsion

Owing to the rapid development of high-energy propellants, the temperature of the solid rocket motor (SRM) combustion chamber has increased, necessitating the development of high-performance polymer matrix composites (PMCs) that meet the requirements of thermal protection systems (TPSs) for use in the ultra-high temperature combustion chamber. ZrO2 is an excellent and potential ultra-high temperature filler, the ablation properties of ZrO2-reinforced PMCs were investigated. ZrO2 reduced the ablation rate of the PMCs by up to 30.6 %. In order to optimize the design of ZrO2 filler, the reinforcement mechanism of ZrO2 on the ablation resistance of the PMCs was explained from the perspective of the properties and chemical reaction in the char layer. The results showed that ZrO2 can increase the residue rate and promote the densification of the char layer. In addition, ZrO2 promoted the formation of ZrC in the char layer, thereby improving the ablation resistance of the ZrO2-reinforced PMCs. The ZrO2-reinforced PMCs are expected to meet the requirements of the TPS used in the ultra-high temperature combustion chamber. Furthermore, the reinforcement mechanism of ZrO2 is conducive to the design of TPSs materials with excellent ablation performance and development of predictive models for the ablation performance of composites.

Predictive performance models in marathon based on half-marathon, age group and pacing behavior

Abstract Objective The main aim of this study was to develop an equation for predicting performance in 42.2 km (MRT) using pacing and packing behavior, age group and previous 21.1 km time as possible explanatory variables. Methods 1571 men and 251 female runners who took part in the Valencia Marathon and Half-Marathon were selected to display the regression models. Stepwise regression analysis showed as explanatory variables for MRT: pacing behavior, age group, and time in 21.1 km. Results The analysis showed four regression models to estimate accurately MRT based principally on athletes previous performance in half-marathon and pacing behavior for men (R2= 0.72–0.88; RMSE= 4:03–8:31 [min:s]). For women, it was suggested a multiple linear regression for estimating MRT (R2 0.95; RSE= 8:06 [min:s]) based on previous performance in half-marathon and pacing behavior. The subsequent concordance analysis showed no significant differences between four of the total regressions with real time in the marathon (p&gt;0.05). Conclusion The present results suggest that even and negative pacing behavior and a better time in 21.1 km, in the previous weeks of the marathon, might accurately predict the MRT. At the same time, nomadic packing behavior was the one that reported the best performance. On the other hand, although the age group variable might partially explain the final performance, it should be included with caution in the final model because of differences in sample distribution, causing an overestimation or underestimation of the final time.

A predictive model for daylight performance based on multimodal generative adversarial networks at the early design stage

In the early building design stage, machine learning-based surrogate models demonstrate significant advantages in rapidly evaluating design performance compared to conventional simulation software. However, due to network limitations, current models struggle to incorporate comprehensive building information, leading to undesired generalizability in adapting to design variations. This study proposes a multimodal Generative Adversarial Network (GAN)-based surrogate model, representing building features with combined multimodal data of images and vectors. Geometric features on the plan and façade are translated into grayscale maps, indicating the potential for daylight reception from each viewpoint, while material properties and weather features are extracted as vectors, such as reflectance, transmittance, sun position and sky conditions. To satisfy the demands of multimodal inputs, the multimodal feature fusion block, the vector-based feature encoding block, and the feature reinforcement block are introduced. These components aim to promote the deep integration of information from different modalities, balance the influences of features across different dimensional scales, and ensure the integrity of the information. The model's validity was verified on the RPLAN database, predicting illumination distribution at the whole-building hierarchy instead of the single-room hierarchy. The results indicate that the multimodal GAN has a Mean Squared Error of 8.129, a Mean Absolute Percentage Error of 0.135, and a Structural Similarity Index of 0.907 on the test set. Meanwhile, this approach saved 73.03% of computing time compared to simulation methods. Boasting strengths in generalizability, speed, and accuracy, the model provides effective support for daylight organization in the early stage of building design, without being confined by variable design scenarios.