Model For Predicting Performance Research Articles

A Comprehensive Discovery Platform for Organophosphorus Ligands for Catalysis.

The design of molecular catalysts typically involves reconciling multiple conflicting property requirements, largely relying on human intuition and local structural searches. However, the vast number of potential catalysts requires pruning of the candidate space by efficient property prediction with quantitative structure-property relationships. Data-driven workflows embedded in a library of potential catalysts can be used to build predictive models for catalyst performance and serve as a blueprint for novel catalyst designs. Herein we introduce kraken, a discovery platform covering monodentate organophosphorus(III) ligands providing comprehensive physicochemical descriptors based on representative conformer ensembles. Using quantum-mechanical methods, we calculated descriptors for 1558 ligands, including commercially available examples, and trained machine learning models to predict properties of over 300000 new ligands. We demonstrate the application of kraken to systematically explore the property space of organophosphorus ligands and how existing data sets in catalysis can be used to accelerate ligand selection during reaction optimization.

Unbiased estimator for the variance of the leave-one-out cross-validation estimator for a Bayesian normal model with fixed variance

When evaluating and comparing models using leave-one-out cross-validation (LOO-CV), the uncertainty of the estimate is typically assessed using the variance of the sampling distribution. Considering the uncertainty is important, as the variability of the estimate can be high in some cases. Previous studies show that no general unbiased variance estimator can be constructed, that would apply for any utility or loss measure and any model. We show that it is possible to construct an unbiased estimator considering a specific predictive performance measure and model. We demonstrate an unbiased sampling distribution variance estimator for the Bayesian normal model with fixed model variance using the expected log pointwise predictive density (elpd) utility score. This example demonstrates that it is possible to obtain improved, problem-specific, unbiased estimators for assessing the uncertainty in LOO-CV estimation.

Modeling and Validation of the Impact of Electric Current and Ambient Temperature on the Thermoelectric Performance of Lithium‐Ion Batteries

A wider electric vehicle penetration strategy demands an improvement in charging time, performance, battery life, and safety, which can be achieved through battery temperature control. A collective impact of load and ambient on the battery needs assessment for thermal management system optimization. Herein, a multiscale multidomain numerical model is utilized on ANSYS Fluent platform. Galvanostatic and constant current discharge tests are conducted for battery modelling and validation. The first of three aspects considered is the cyclic relationship between the thermal parameters. Low ambient temperature of 0 °C at 1 C discharge rate increases the heat generation rate to 50 kW m−3 which then dips to 40 kW m−3 due to rise in battery temperature. The second is the inter‐relationship among the thermal and the electrical parameters. A faster rise in battery temperature at low ambient temperature causes an elevation in the voltage and hence the electrical performance. The third is the collective impact of thermal and electrical parameters. At 0 °C ambient temperature, an increase in discharge rate from 1 to 1.5 C increases the battery capacity by at least 6%. With the establishment of the predictive performance model of the battery, this study is expected to provide a guideline for optimizing the thermal management systems.

Development and application of an in-situ indentation testing system for the prediction of tunnel boring machine performance

Prediction of the tunnel boring machine (TBM) performance is critically important for project scheduling and cost estimation. Various models have been proposed to estimate the TBM performance. However, most of these models used the parameters of rock specimens acquired from the laboratory tests, which differed from the in-situ conditions. It is difficult to apply these models to the engineering practice. In the present study, to overcome these limitations, an in-situ indentation testing system for the force-penetration response was proposed and applied to a TBM-excavated tunnel. A database that integrates 33 groups of in-situ indentation testing (235 tests in total) and the corresponding TBM operating parameters was established, including 12 indentation indices extracted from the force-penetration curve and 6 operating parameters recorded by the TBM. A series of correlation analysis was carried out to investigate the relationships between these indices/parameters. The analysis indicates that the 12 indentation indices are highly or moderately correlated with each other. A similar phenomenon was found in the TBM operating parameters. Consequently, four indentation indices and one operating parameter (field penetration index, FPI) were selected to establish the predictive models. Predictive models of the TBM performance (FPI) with high correlation coefficient ( r = 0.960) were proposed using the multiple indices obtained from the in-situ indentation testing.

Machine learning prediction of cognition from functional connectivity: Are feature weights reliable?

Cognitive performance can be predicted from an individual's functional brain connectivity with modest accuracy using machine learning approaches. As yet, however, predictive models have arguably yielded limited insight into the neurobiological processes supporting cognition. To do so, feature selection and feature weight estimation need to be reliable to ensure that important connections and circuits with high predictive utility can be reliably identified. We comprehensively investigate feature weight test-retest reliability for various predictive models of cognitive performance built from resting-state functional connectivity networks in healthy young adults (n=400). Despite achieving modest prediction accuracies (r=0.2–0.4), we find that feature weight reliability is generally poor for all predictive models (ICC< 0.3), and significantly poorer than predictive models for overt biological attributes such as sex (ICC≈0.5). Larger sample sizes (n=800), the Haufe transformation, non-sparse feature selection/regularization and smaller feature spaces marginally improve reliability (ICC< 0.4). We elucidate a tradeoff between feature weight reliability and prediction accuracy and find that univariate statistics are marginally more reliable than feature weights from predictive models. Finally, we show that measuring agreement in feature weights between cross-validation folds provides inflated estimates of feature weight reliability. We thus recommend for reliability to be estimated out-of-sample, if possible. We argue that rebalancing focus from prediction accuracy to model reliability may facilitate mechanistic understanding of cognition with machine learning approaches.

Assessing and Augmenting Predictive Models for Hospital Readmissions With Novel Variables in an Urban Safety-net Population.

The performance of existing predictive models of readmissions, such as the LACE, LACE+, and Epic models, is not established in urban safety-net populations. We assessed previously validated predictive models of readmission performance in a socially complex, urban safety-net population, and if augmentation with additional variables such as the Area Deprivation Index, mental health diagnoses, and housing access improves prediction. Through the addition of new variables, we introduce the LACE-social determinants of health (SDH) model. This retrospective cohort study included adult admissions from July 1, 2016, to June 30, 2018, at a single urban safety-net health system, assessing the performance of the LACE, LACE+, and Epic models in predicting 30-day, unplanned rehospitalization. The LACE-SDH development is presented through logistic regression. Predictive model performance was compared using C-statistics. A total of 16,540 patients met the inclusion criteria. Within the validation cohort (n=8314), the Epic model performed the best (C-statistic=0.71, P<0.05), compared with LACE-SDH (0.67), LACE (0.65), and LACE+ (0.61). The variables most associated with readmissions were (odds ratio, 95% confidence interval) against medical advice discharge (3.19, 2.28-4.45), mental health diagnosis (2.06, 1.72-2.47), and health care utilization (1.94, 1.47-2.55). The Epic model performed the best in our sample but requires the use of the Epic Electronic Health Record. The LACE-SDH performed significantly better than the LACE and LACE+ models when applied to a safety-net population, demonstrating the importance of accounting for socioeconomic stressors, mental health, and health care utilization in assessing readmission risk in urban safety-net patients.

Predictive models for daylight performance of general floorplans based on CNN and GAN: A proof-of-concept study

A daylight performance evaluation at the early design stage is essential for a building morphology design and optimization, having a tremendous influence on energy consumption and indoor environments. Considering the complicated input parameters, time and computational cost of the simulation tools, although proxy models based on various machine learning algorithms have been developed, they are limited to certain building forms described based on the selected parameters. In this study, proxy models of a daylight simulation for general floorplans are proposed based on convolutional neural network (CNN) and generative adversarial network (GAN). ResNet (CNN) and pix2pix (GAN) are applied to predict static and annual daylight metrics (uniformity, mean lux, success rate, sDA, and UDI) and illuminance distribution in space, respectively. Geometry information is embedded in the image structure, and a grid-based simulation are conducted as the ground truth. Two datasets composed of real floorplan cases and parametric rooms were tested in the experiments. ResNet obtains the best R2 of 0.959 and an MSE of 0.008 on the real case dataset for daylight uniformity. In addition, pix2pix generates visualization results close to the simulation with an SSIM of 0.90 in the test set within a period of 1 s and provides real-time intuitive feedback for designers. The results show the possibility of using deep neural networks to extract features from general building forms and build predictive models, which can be integrated into automatic form-finding and design optimization.

The Relationship Between High-intensity Functional Training Performance And Laboratory-based Measurements Of Fitness

To date, research has examined the physiological determinants of performance in standardized High-Intensity Functional Training (HIFT) workouts but not without the influence of HIFT familiarity. PURPOSE: To examine the predictive value of aerobic fitness, body composition, and total body strength on performance of two standardized HIFT workouts in HIFT-naïve participants. METHODS: Twenty-two recreationally trained individuals (males = 13, females = 9) underwent assessments of peak oxygen consumption (VO2 peak), ventilatory thresholds, body composition, and one repetition maximum tests for the back squat, deadlift, and overhead press in which the sum equaled the total strength. Participants also performed two standardized HIFT workouts: a scaled version of the 19.1 (as many sets of 19 wall balls followed by 19 calories on rower within 15 minutes) and a modified Fran (sets of 21, 15, and 9 repetitions of thrusters and burpees) to determine scores based on total repetitions completed and time-to-completion, respectively. Simple Pearson’s r correlations were used to determine the relationships between HIFT performance variables (19.1 and modified Fran) and the independent variables. A forward stepwise multiple linear regression analysis was performed and significant variables that survived the regression analysis were used to create a predictive model of HIFT performance. RESULTS: Absolute VO2 peak was a significant predictor of 19.1 performance, explaining 39% of its variance (adjusted R2 = 0.39, p = 0.002). For modified Fran, total strength was a significant predictor and explained 33% of the variance in performance (adjusted R2 = 0.33, p = 0.005). CONCLUSION: These results suggest, without any influence of HIFT familiarity or experience, that performance in these two HIFT workouts could be predicted by distinct laboratory-based measurements of fitness.

From intuition to intelligence: a text mining–based approach for movies' green-lighting process

PurposeThe purpose of this paper is to develop a predictive model for box office performance based on the textual information in movie scripts in the green-lighting process of movie production.Design/methodology/approachThe authors use Latent Dirichlet Allocation to determine the hidden textual structure in movie scripts by extracting topic probabilities as predictors for classification. The extracted topic probabilities are used as inputs for the predictive model for the box office performance. For the predictive model, the authors utilize a variety of classification algorithms such as logistic classification, decision trees, random forests, k-nearest neighbor algorithms, support vector machines and artificial neural networks, and compare their relative performances in predicting movies' market performance.FindingsThis approach for extracting textual information from movie scripts produces a valuable typology for movies. Moreover, our modeling approach has significant power to predict movie scripts' profitability. It provides a superior prediction performance compared to previous benchmarks, such as that of Eliashberg et al. (2007).Research limitations/implicationsThis work contributes to literature on predicting the box office performance in the green-lighting process and literature regarding suggesting models for the idea screening stage in the new product development process. Besides, this is one of the few studies that use movie script data to predict movies' financial performance by proposing an approach to integrate text mining models and machine learning algorithms with movie experts' intuition.Practical implicationsFirst, the authors’ approach can significantly reduce the financial risk associated with movie production decisions before the pre-production stage. Second, this paper proposes an approach that is applicable at a very early stage of new product development, such as the idea screening stage. The authors also introduce an online-based movie scenario database system that can help movie studios make more systematic and profitable decisions in the green-lighting process. Third, this approach can help movie studios estimate movie scripts' financial value.Originality/valueThis study is one of the few studies to forecast market performance in the green-lighting process.

Pulse Voltammetry Analysis of Li-Ion Battery Degradation

Pulse voltammetry is a class of electroanalytical techniques which involve applying voltage pulses to understand the redox reaction in an electrochemical system. The advantages of such a technique is high sensitivity of pulse voltammograms towards kinetic and thermodynamic parameters. Here, a physics-based model is developed to simulate the Li-ion battery physics and the resultant pulse voltammogram for generic voltage pulses. We develop an in-situ diagnostic tool by combining it with a parameter estimation technique to accurately, rapidly and quantitatively capture the key parameters. A combined experimental and computational-modelling approach allows us to parametrize the state of Li-ion battery system. Degradation reactions such as solid electrolyte interphase growth, active material slippage, and loss have been considered to capture the pertinent degradation modes responsible for capacity fade under different experimental conditions. This method can be extended to integrate with battery management systems for state of health estimation and monitoring.The objective of this effort is to quantitatively characterize battery properties in order to develop accurate physics-based models suitable for model-based development of fast charging protocols and accurate state-of-charge and state-of-health prediction. Pulse voltammetry analysis software and battery electrode characterization approaches provide the ability to determine quantitative values, with known confidence intervals, for input parameters of physics-based models describing battery performance. Pulse voltammetry characterization, in conjunction with physical analysis and complementary techniques such as galvanostatic intermittent titration (GITT) and electrochemical impedance spectroscopy (EIS) will be utilized to analyze the battery state. These quantitative methods enable the development and parameterization of predictive models for both battery performance and undesired processes, such as lithium plating, that limit charging rates and cycle life of lithium-ion batteries.This presentation will focus on DPV data gathered periodically during cell cycling to identify the progression of capacity loss and degradation. The DPV analysis approach, which utilizes the experimental data to estimate kinetic and thermodynamic parameters of the model for battery performance, will be presented. The experimental characterization results will include example DPV measurements as a function of state-of-charge, cycling history, and operational temperature. Examples of data analysis to estimate model parameters with statistical confidence measures, and interpretation of the change in voltammogram response with cycling will be discussed. Figure 1. Fitting model results with experimental differential pulse voltammetry data for parameterization Figure 1

Classification of Imbalanced Data Stream: Techniques and Challenges

As the number of generated data increases every day, this has brought the importance of data mining and knowledge extraction. In traditional data mining, offline status can be used for knowledge extraction. Nevertheless, dealing with stream data mining is different due to continuously arriving data that can be processed at a single scan besides the appearance of concept drift. As the pre-processing stage is critical in knowledge extraction, imbalanced stream data gain significant popularity in the last few years among researchers. Many real-world applications suffer from class imbalance including medical, business, fraud detection and etc. Learning from the supervised model includes classes whether it is binary- or multi-classes. These classes are often imbalance where it is divided into the majority (negative) class and minority (positive) class, which can cause a bias toward the majority class that leads to skew in predictive performance models. Handles imbalance streaming data is mandatory for more accurate and reliable learning models. In this paper, we will present an overview of data stream mining and its tools. Besides, summarize the problem of class imbalance and its different approaches. In addition, researchers will present the popular evaluation metrics and challenges prone from imbalanced streaming data.

Portability of Predictive Academic Performance Models: An Empirical Sensitivity Analysis

The portability of predictive models of academic performance has been widely studied in the field of learning platforms, but there are few studies in which the results of previous evaluations are used as factors. The aim of this work was to analyze portability precisely in this context, where preceding performance is used as a key predictor. Through a study designed to control the main confounding factors, the results of 170 students evaluated over two academic years were analyzed, developing various predictive models for a base group (BG) of 39 students. After the four best models were selected, they were validated using different statistical techniques. Finally, these models were applied to the remaining groups, controlling the number of different factors with respect to the BG. The results show that the models’ performance varies consistently with what was expected: as they move away from the BG (fewer common characteristics), the specificity of the four models tends to decrease.

Influence of Two Mass Variables on Inertia Cone Crusher Performance and Optimization of Dynamic Balance

Inertia cone crushers are widely used in complex ore mineral processing. The two mass variables (fixed cone mass and moving cone mass) affect the dynamic performance of the inertia cone crusher. Particularly the operative crushing force of the moving cone and the amplitude of the fixed cone are affected, and thus the energy consumption of the crusher. In this paper, the process of crushing steel slag is taken as a specific research object, to analyze the influence of two mass variables on the inertia cone crusher performance. A real-time dynamic model based on the multi-body dynamic (MBD) and the discrete element method (DEM) is established. Furthermore, the influence of the fixed cone mass and moving cone mass on the operative crushing force, amplitude and average power draw are explored by the design of simulation experiments. The predictive regression models of inertia cone crusher performance are obtained using response surface methodology (RSM). After increasing the fixed cone mass, the optimized amplitude, average power and moving cone mass are decreased by 37.1%, 33.1% and 10%, respectively, compared to without the adjustment. Finally, a more effective dynamic balancing mechanism of inertia cone crusher is achieved, which can utilize the kinetic energy of a balancer, and minimize the mass of the fixed and moving cone. The fixed cone mass and moving cone mass of a balancing crusher are decreased by 78.9% and 22.8%, respectively, compared to without the balancing mechanism.

Decision tree learning through a Predictive Model for Student Academic Performance in Intelligent M-Learning environments

Abstract In the area of machine learning and data science, decision tree learning is considered as one of the most popular classification techniques. Therefore, a decision tree algorithm generates a classification and predictive model, which is simple to understand and interpret, easy to display graphically, and capable to handle both numerical and categorical data. The intelligent m-learning systems, enjoy recently an explosive growth of interest, for more effective education and adaptive learning tailored to each student's learning abilities. The goal of this paper is to further improve personalization in student academic performance, that includes dynamic tests with a predictive model. One major objective of this research is to create adaptive dynamic tests for assessing student academic performance, while constantly comparing the results of the assessment which exhibits the individual student profile, with the results of the decision tree's algorithm which formulates a predictive model for students' knowledge level, according to the weights of the decision tree.

Empowering Organisations to Gain From Uncertainty: a Conceptualisation of Antifragility Through Leveraging Organisational Routines in Uncertain Environments

Economic shifts, disruptive innovations, and competitive rivalries continuously reshape the operating environment of organisations. Such uncertainty impacts organisations and raises significant challenges. While many organisations tend to respond to uncertainty by adopting loss minimisation strategies, others see uncertainty as an opportunity to achieve gains. The latter view is exemplified in Taleb’s (2012) concept of ‘antifragility’, a property of systems that gain when exposed to uncertainty. For organisations, the challenge lies in the identification and execution of fundamental artefacts to accomplish work to achieve antifragile outcomes. One such artefact is the organisational routine; repeatable, regular patterns of behaviour and actions that influence performance. This paper conceptualises the intersection between antifragility, uncertainty management, and organisational routines literatures to identify four routine archetypes that can guide actions that contribute to organisational antifragility. Theoretically, this paper identifies how these archetypes arise from the interplay between temporal action (as tendencies towards proactive or reactive action) and risk mitigation strategies (as preference towards redundancies or flexibilities). Developed insights bring forth a foundation for predictive models of performance, and guidance for organisations aiming to thrive, rather than just survive, in uncertain environments. This paper concludes with the identification of further research avenues.

Multiclass Prediction Model for Student Grade Prediction Using Machine Learning

Today, predictive analytics applications became an urgent desire in higher educational institutions. Predictive analytics used advanced analytics that encompasses machine learning implementation to derive high-quality performance and meaningful information for all education levels. Mostly know that student grade is one of the key performance indicators that can help educators monitor their academic performance. During the past decade, researchers have proposed many variants of machine learning techniques in education domains. However, there are severe challenges in handling imbalanced datasets for enhancing the performance of predicting student grades. Therefore, this paper presents a comprehensive analysis of machine learning techniques to predict the final student grades in the first semester courses by improving the performance of predictive accuracy. Two modules will be highlighted in this paper. First, we compare the accuracy performance of six well-known machine learning techniques namely Decision Tree (J48), Support Vector Machine (SVM), Naïve Bayes (NB), K-Nearest Neighbor (kNN), Logistic Regression (LR) and Random Forest (RF) using 1282 real student's course grade dataset. Second, we proposed a multiclass prediction model to reduce the overfitting and misclassification results caused by imbalanced multi-classification based on oversampling Synthetic Minority Oversampling Technique (SMOTE) with two features selection methods. The obtained results show that the proposed model integrates with RF give significant improvement with the highest f-measure of 99.5%. This proposed model indicates the comparable and promising results that can enhance the prediction performance model for imbalanced multi-classification for student grade prediction.

A Predictive Model for the Long-Term Electrical Performance of a Leadless Transcatheter Pacemaker

This study sought to formulate a predictive model for describing the long-term electrical performance of Micra (Medtronic, Mounds View, Minnesota). The Micra leadless pacemaker is an alternative ventricular pacing option that avoids the pitfalls of transvenous leads. However, well-defined metrics to predict the long-term electrical performance of the device are lacking. We identified all patients who underwent successful Micra implantation enrolled in the investigational device exemption study, continued access study, or post-approval registry with complete 1-year post-implantation data or system revision due to elevated thresholds (N=1,843). The analysis endpoint was an elevated pacing capture threshold (PCT) at≥12months post-implantation, defined as≥2.0V at 0.24ms or an increase of≥1.5V from implantation or need for system revision due to elevated thresholds at≤12months post-implantation. We evaluated for univariate and multivariate associations between patient and device characteristics at implantation and for elevated thresholds at 12months. Among the total cohort, 75 patients (4.1%) had elevated thresholds at 12months; of these, 13 required system revisions. Predictors associated with elevated thresholds in univariate analysis included the total number of deployments (excluded from the multivariable model), impedance and PCT at implantation, male sex, history of diabetes, and ischemic cardiomyopathy. Multivariable regression modeling found that male sex, history of diabetes, implantation PCT of≥2 V, and impedance of<800 Ω were independent predictors of elevated PCT at 12months (all p<0.05). A history of diabetes, male sex, elevated PCT, and low impedance at implantation were independentpredictors of elevated thresholds at 12months. These metrics represent the foundation of a simple tooltoaid in procedural decision making.

An Exploration of Factors Linked to Academic Performance in PISA 2018 Through Data Mining Techniques.

International large-scale assessments, such as PISA, provide structured and static data. However, due to its extensive databases, several researchers place it as a reference in Big Data in Education. With the goal of exploring which factors at country, school and student level have a higher relevance in predicting student performance, this paper proposes an Educational Data Mining approach to detect and analyze factors linked to academic performance. To this end, we conducted a secondary data analysis and built decision trees (C4.5 algorithm) to obtain a predictive model of school performance. Specifically, we selected as predictor variables a set of socioeconomic, process and outcome variables from PISA 2018 and other sources (World Bank, 2020). Since the unit of analysis were schools from all the countries included in PISA 2018 (n = 21,903), student and teacher predictor variables were imputed to the school database. Based on the available student performance scores in Reading, Math, and Science, we applied k-means clustering to obtain a categorized (three categories) target variable of global school performance. Results show the existence of two main branches in the decision tree, split according to the schools’ mean socioeconomic status (SES). While performance in high-SES schools is influenced by educational factors such as metacognitive strategies or achievement motivation, performance in low-SES schools is affected in greater measure by country-level socioeconomic indicators such as GDP, and individual educational indicators are relegated to a secondary level. Since these evidences are in line and delve into previous research, this work concludes by analyzing its potential contribution to support the decision making processes regarding educational policies.

Predictive Performance Models in Long-Distance Runners: A Narrative Review.

Physiological variables such as maximal oxygen uptake (VO2max), velocity at maximal oxygen uptake (vVO2max), running economy (RE) and changes in lactate levels are considered the main factors determining performance in long-distance races. The aim of this review was to present the mathematical models available in the literature to estimate performance in the 5000 m, 10,000 m, half-marathon and marathon events. Eighty-eight articles were identified, selections were made based on the inclusion criteria and the full text of the articles were obtained. The articles were reviewed and categorized according to demographic, anthropometric, exercise physiology and field test variables were also included by athletic specialty. A total of 58 studies were included, from 1983 to the present, distributed in the following categories: 12 in the 5000 m, 13 in the 10,000 m, 12 in the half-marathon and 21 in the marathon. A total of 136 independent variables associated with performance in long-distance races were considered, 43.4% of which pertained to variables derived from the evaluation of aerobic metabolism, 26.5% to variables associated with training load and 20.6% to anthropometric variables, body composition and somatotype components. The most closely associated variables in the prediction models for the half and full marathon specialties were the variables obtained from the laboratory tests (VO2max, vVO2max), training variables (training pace, training load) and anthropometric variables (fat mass, skinfolds). A large gap exists in predicting time in long-distance races, based on field tests. Physiological effort assessments are almost exclusive to shorter specialties (5000 m and 10,000 m). The predictor variables of the half-marathon are mainly anthropometric, but with moderate coefficients of determination. The variables of note in the marathon category are fundamentally those associated with training and those derived from physiological evaluation and anthropometric parameters.

A Faster Path to Solar Water Splitting

Finding new wide-bandgap light absorbers that are stable in aqueous solutions is a long-standing challenge in photoelectrochemical water splitting research. Two papers in this issue describe recent advances in high throughput experimentation that may accelerate the discovery of suitable materials.