Statistical Learning Models Research Articles

A comparison of statistical and machine learning models for spatio-temporal prediction of ambient air pollutant concentrations in Scotland

AbstractThe spatio-temporal prediction of air pollutant concentrations is vital for assessing regulatory compliance and for producing exposure estimates in epidemiological studies. Numerous approaches have been utilised for making such predictions, including land use regression models, additive models, spatio-temporal smoothing models and machine learning prediction algorithms. However, relatively few studies have compared the predictive performance of these models thoroughly, which is one of the novel contributions of this paper. For the specific challenge of predicting monthly average concentrations of NO$$_{2}$$ 2 , PM$$_{10}$$ 10 and $$\hbox {PM}_{2.5}$$ PM 2.5 in Scotland, we find that random forests typically outperform (or are as good as) more traditional statistical prediction approaches. Additionally, we utilise the best performing model to provide a new data resource, namely, predictions of monthly average concentrations (with uncertainty quantification) of the above pollutants on a regular 1 km$$^{2}$$ 2 grid for all of Scotland between 2016 and 2020.

Rotating Bending Fatigue Durability of Wrought 304 Stainless Steel: A Comparative Study of Surface-Treated vs. Untreated Specimen

The rotating bending fatigue test, which is a common durability test for industrial components, was employed in this study to examine the impact of the surface treatment procedure on the material's performance under room temperature. The studied material in this research is a wrought stainless steel (SS) 304 alloy that has undergone a specialized surface treatment to enhance the fatigue properties of this unique material. The fatigue endurance of untreated materials was 405 MPa (65% of untreated specimen’s yield strength) on 5 million runout cycles, while the surface-treated samples exhibited 545 MPa (85% of treated specimen’s yield strength), a significantly greater value compared to the untreated ones (even up to 10 million runout cycles). The experimental results demonstrated that the surface-treated samples exhibited no signs of cracking. Instead, all failures were attributed to plastic bending, particularly at higher stress levels. Furthermore, the samples displayed a limited range of stress levels, spanning from endurance to plastic bending conditions. This finding suggests that the hypothesis of an increase in stiffness in the surface layer is more pronounced than in the material interior. It also indicates that the surface layer material was not cracked under the maximum stress level, while the layer beneath it experienced plastic deformation. A comparison has been made between the Basquin equation and a statistical machine learning model (for a limited amount of data) to incorporate the impact of surface treatment on predicting fatigue lifetime, and the random forest regressor model would be able to predict the lifetime with the accuracy of R2= 87% which was much higher than the classical model.

Refined selection of individuals for preventive cardiovascular disease treatment with a transformer-based risk model

Abstract Background Although statistical models have been commonly used to identify patients at risk of cardiovascular disease (CVD) for preventive therapy, these models tend to over-recommend therapy [1]. Moreover, in populations with pre-existing diseases, the current approach is to indiscriminately treat all individuals, as modelling in this context is currently inadequate. Aims To develop and validate a novel deep learning model, the Transformer-based Risk assessment survival model (TRisk) model, for prediction of 10-year risk of CVD in both the general population and individuals with diabetes. Methods An open cohort of 3 million adults aged 25 to 84 years free of CVD and statin use was identified using linked primary and secondary electronic health records from 291 and 98 general practices in England, which were used for TRisk model development and external validation, respectively. Comparison against two commonly recommended statistical models (QRISK3 [2] and SCORE2 [3]) and a deep learning derivation of QRISK3 was conducted. Additional analyses compared discriminatory performance in other age groups, by sex, and across categories of socioeconomic status. Similar analyses were conducted in a subgroup of 59 thousand patients with diabetes. Results TRisk demonstrated superior discrimination (C index in the general population: 0.910; 95% confidence interval [CI]: 0.906 to 0.913). TRisk’s performance was found to be less sensitive to population age range than the benchmark models and outperformed other models also in all stratified analyses. All models were overall well-calibrated. In decision curve analyses, TRisk demonstrated greater net benefit than benchmark models across the range of relevant thresholds. In impact analyses, at the widely recommended 10% risk threshold and the higher 15% threshold for the general population, TRisk reduced both the total number of patients classified at high risk (by 27% and 39% respectively) and the number of false negatives as compared with currently recommended strategies. TRisk similarly outperformed other models in patients with diabetes. Compared with the widely recommended treat-all policy approach for patients with diabetes, TRisk at a 10% risk threshold would lead to deselection of 24% of individuals with a small fraction of false negatives (0.2% of cohort). Conclusion TRisk enabled a more targeted selection of individuals at risk of CVD compared to benchmark statistical and deep learning models, in both the general population and patients with diabetes. Incorporation of TRisk into routine clinical care would allow a reduction in the number of treatment-eligible patients in general population and subgroup with diabetes by approximately one-third and one-quarter respectively while preventing at least as many events as with currently adopted approaches.Models' discrimination on both cohortsImpact analyses for both cohorts

Fixing imbalanced binary classification: An asymmetric Bayesian learning approach.

Most statistical and machine learning models used for binary data modeling and classification assume that the data are balanced. However, this assumption can lead to poor predictive performance and bias in parameter estimation when there is an imbalance in the data due to the threshold election for the binary classification. To address this challenge, several authors suggest using asymmetric link functions in binary regression, instead of the traditional symmetric functions such as logit or probit, aiming to highlight characteristics that would help the classification task. Therefore, this study aims to introduce new classification functions based on the Lomax distribution (and its variations; including power and reverse versions). The proposed Bayesian functions have proven asymmetry and were implemented in a Stan program into the R workflow. Additionally, these functions showed promising results in real-world data applications, outperforming classical link functions in terms of metrics. For instance, in the first example, comparing the reverse power double Lomax (RPDLomax) with the logit link showed that, regardless of the data imbalance, the RPDLomax model assigns effectively lower mean posterior predictive probabilities to failure and higher probabilities to success (21.4% and 63.7%, respectively), unlike Logistic regression, which does not clearly distinguish between the mean posterior predictive probabilities for these two classes (36.0% and 39.5% for failure and success, respectively). That is, the proposed asymmetric Lomax approach is a competitive model for differentiating binary data classification in imbalanced tasks against the Logistic approach.

Machine learning-based prediction models unleash the enhanced production of fucoxanthin in Isochrysis galbana.

The marine microalga Isochrysis galbana is prolific producer of fucoxanthin, which is a xanthophyll carotenoid with substantial global market value boasting extensive applications in the food, nutraceutical, pharmaceutical, and cosmetic industries. This study presented a novel integrated experimental approach coupled with machine learning (ML) models to predict the fucoxanthin content in I. galbana by altering the type and concentration of phytohormone supplementation, thus overcoming the multiple methodological limitations of conventional fucoxanthin quantification. A novel integrated experimental approach was developed, analyzing the effect of varying phytohormone types and concentrations on fucoxanthin production in I. galbana. Morphological analysis was conducted to assess changes in microalgal structure, while growth rate and fucoxanthin yield correlations were explored using statistical analysis and machine learning models. Several ML models were employed to predict fucoxanthin content, with and without hormone descriptors as variables. The findings revealed that the Random Forest (RF) model was highly significant with a high of 0.809 and of 0.776 when hormone descriptors were excluded, and the inclusion of hormone descriptors further improved prediction accuracy to of 0.839, making it a useful tool for predicting the fucoxanthin yield. The model that fitted the experimental data indicated methyl jasmonate (0.2 mg/L) as an effective phytohormone. The combined experimental and ML approach demonstrated rapid, reliable, and cost-efficient prediction of fucoxanthin yield. This study highlights the potential of machine learning models, particularly Random Forest, to optimize parameters influencing microalgal growth and fucoxanthin production. This approach offers a more efficient alternative to conventional methods, providing valuable insights into improving fucoxanthin production in microalgal cultivation. The findings suggest that leveraging diverse ML models can enhance the predictability and efficiency of fucoxanthin production, making it a promising tool for industrial applications.

Predictive modeling of biomedical temporal data in healthcare applications: review and future directions.

Predictive modeling of clinical time series data is challenging due to various factors. One such difficulty is the existence of missing values, which leads to irregular data. Another challenge is capturing correlations across multiple dimensions in order to achieve accurate predictions. Additionally, it is essential to take into account the temporal structure, which includes both short-term and long-term recurrent patterns, to gain a comprehensive understanding of disease progression and to make accurate predictions for personalized healthcare. In critical situations, models that can make multi-step ahead predictions are essential for early detection. This review emphasizes the need for forecasting models that can effectively address the aforementioned challenges. The selection of models must also take into account the data-related constraints during the modeling process. Time series models can be divided into statistical, machine learning, and deep learning models. This review concentrates on the main models within these categories, discussing their capability to tackle the mentioned challenges. Furthermore, this paper provides a brief overview of a technique aimed at mitigating the limitations of a specific model to enhance its suitability for clinical prediction. It also explores ensemble forecasting methods designed to merge the strengths of various models while reducing their respective weaknesses, and finally discusses hierarchical models. Apart from the technical details provided in this document, there are certain aspects in predictive modeling research that have arisen as possible obstacles in implementing models using biomedical data. These obstacles are discussed leading to the future prospects of model building with artificial intelligence in healthcare domain.

A parallel recursive framework for modelling time series

Abstract Time series modelling is of significance to several scientific fields. Several approaches based on statistics, machine learning or combinations have been utilized. In order to model and forecast time series a novel parallel framework based on recursive pseudoinverse matrices is proposed. This framework enables the design of arbitrary statistical and machine learning models, adaptively, from a set of potential basis functions. This unification enables compact definition of existing and new models as well as easy implementation for new massively parallel architectures. The choice of appropriate basis functions is analysed and the fitting accuracy, termination criteria and model update operations are presented. A block variant for multivariate time series is also proposed. Parallel GPU implementation and performance optimization of the framework are provided, based on mixed precision arithmetic and matrix operations. The use of different basis functions is showcased with respect to various model univariate and multivariate time series for applications such as regression, frequency estimation and automatic trend detection. Discussions on limitations and future directions of research are also provided.

Assessment and prediction of Water Quality Index (WQI) by seasonal key water parameters in a coastal city: application of machine learning models.

The Water Quality Index (WQI) provides comprehensive assessments in river systems; however, its calculation involves numerous water quality parameters, costly in sample collection and laboratory analysis. The study aimed to determine key water parameters and the most reliable models, considering seasonal variations in the water environment, to maximize the precision of WQI prediction by a minimal set of water parameters. Ten statistical or machine learning models were developed to predict the WQI over four seasons using water quality dataset collected in a coastal city adjacent to the Yellow Sea in China, based on which the key water parameters were identified and the variations were assessed by the Seasonal-Trend decomposition procedure based on Loess (STL). Results indicated that model performance generally improved with adding more input variables except Self-Organizing Map (SOM). Tree-based ensemble methods like Extreme Gradient Boosting (XGB) and Random Forest (RF) demonstrated the highest accuracy, particularly in winter. Nutrients (Ammonia Nitrogen (AN) and Total Phosphorus (TP)), Dissolved Oxygen (DO), and turbidity were determined as key water parameters, based on which, the prediction accuracy for Medium and Low grades was perfect while it was over 80% for the Good grade in spring and winter and dropped to around 70% in summer and autumn. Nutrient concentrations were higher at inland stations; however, it worsened at coastal stations, especially in summer. The study underscores the importance of reliable WQI prediction models in water quality assessment, especially when data is limited, which are crucial for managing water resources effectively.

A probabilistic framework for identifying anomalies in urban air quality data.

Just as the value of crude oil is unlocked through refining, the true potential of air quality data is realized through systematic processing, analysis, and application. This refined data is critical for making informed decisions that may protect health and the environment. Perhaps ground-based air quality monitoring data often face quality control issues, notably outliers. The outliers in air quality data are reported as error and event-based. The error-based outliers are due to instrument failure, self-calibration, sensor drift over time, and the event based focused on the sudden change in meteorological conditions. The event-based outliers are meaningful while error-based outliers are noise that needs to be eliminated and replaced post-detection. In this study, we address error-based outlier detection in air quality data, particularly targeting particulate pollutants (PM2.5 and PM10) across various monitoring sites in Delhi. Our research specifically examines data from sites with less than 5% missing values and identifies four distinct types of error-based outliers: extreme values due to measurement errors, consecutive constant readings and low variance due to instrument malfunction, periodic outliers from self-calibration exceptions, and anomalies in the PM2.5/PM10 ratio indicative of issues with the instruments' dryer unit. We developed a robust methodology for outlier detection by fitting a non-linear filter to the data, calculating residuals between observed and predicted values, and then assessing these residuals using a standardized Z-score to determine their probability. Outliers are flagged based on a probability threshold established through sensitivity testing. This approach helps distinguish normal data points from suspicious ones, ensuring the refined quality of data necessary for accurate air quality modeling. This method is essential for improving the reliability of statistical and machine learning models that depend on high-quality environmental data.

Real-time traffic conflict prediction at signalized intersections using vehicle trajectory data and deep learning

Real-time conflict prediction at signalized intersections is crucial for urban road safety management. This study developed a real-time conflict prediction framework for signalized intersections using video data real-time recognition technology and deep learning techniques, incorporating lane-level information and feature interactions. The modeling framework consists of three stages: real-time video data extraction and processing, the development of a Deep and Cross Network (DCN)-based real-time traffic conflict prediction model, and conflict-driven factor interpretability analysis through SHapley Additive exPlanations (SHAP). In the first stage, an efficient automated trajectory extraction system is designed to obtain vehicle trajectories in real time for dynamic traffic parameters and conflict frequency identification. In the second stage, a DCN model is developed to construct the relationships between dynamic traffic parameters, including their interactions, and traffic conflicts. In the third stage, SHAP explores the impact mechanisms of different dynamic traffic parameters on traffic conflicts. The model’s predictive performance and interpretability are evaluated using intersection video data from Changsha City, China. The results show that: (1) In real-time traffic conflict prediction at signalized intersections across different modified time-to-conflict thresholds (1.5 s and 3.0 s), the DCN model consistently outperformed statistical and machine learning models. (2) High traffic flow on main and secondary roads at signalized intersections significantly increases the complexity and frequency of conflicts, with varying sensitivity depending on the interaction of traffic flow, speed, and platoon length. (3) The proposed framework provides a safety measurement standard for data-driven road safety management methods.

Incorporating multi-source uncertainties in fast building wall thermal resistance estimation through physics-based and multi-fidelity statistical learning models

Meta-modeling and Bayesian inversion technique are proposed for fast and accurate in-situ estimation of the total thermal resistance (RTot) of walls using non-intrusive wall instrumentation. Various sources of uncertainties are taken into account to provide an enhanced credible interval for the thermal resistance estimate. In the considered protocol, a small zone of the internal surface of the wall is excited by a prototype to get faster in situ estimation and to limit the influence of external weather conditions. To be independent of the heat transfer coefficients and of the misknown layer thicknesses, which are difficult to estimate, we use herein the measurements of internal and external surface temperatures as boundary conditions in a thermal direct problem reformulated. A meta-model of the thermal model is created based on a statistical multi-fidelity approach with two levels of fidelity, Resistance-Capacitance (RC) and 1D models, to achieve the Bayesian estimation of the total thermal resistance in a reasonable computation time. The RTot estimation method is applied to realistic internal insulated walls (IIW), from poorly to highly insulated, under different weather conditions. Several numerical tests are carried out and credible intervals are provided to study the importance of the wall initial condition, the excitation time and the instrumentation. Finally, an experimental application is conducted using real measurements on an internal insulated wall (IIW) in Nancy (France). The obtained experimental results show that, in a short excitation time (10h) with a reduced instrumentation, the proposed method accurately estimates the minimum wall thermal resistance. For a standard wall of about 4 m2.K/W, a relevant estimate of RTot with a relative deviation less than 10% can be achieved by using a polynomial initial temperature profile proposed in this study, sufficient measurements and an excitation time of 3 days. This study thus offers prospects for improved energy assessment of buildings before and after renovation.

Evaluation of statistical and machine learning models using satellite data to estimate aboveground biomass: A study in Vietnam Tropical Forests

The combination of machine learning models with satellite imagery is becoming a popular data-modeling tool for biomass prediction, supporting land cover management. This study aims to select the most suitable model to estimate tropical forest aboveground biomass in Vietnam, helping to manage and monitor changes in biomass at regional and local scales. The study identified the optimal model for estimating forest aboveground biomass and minimizing the number of input variables while achieving satisfactory model performance. A total of 59 input variables, including topography, texture features, and vegetation indices, from satellite data were used in four non-parametric algorithms and a conventional parametric model, Artificial Neural Networks (ANN), Support Vector Machine (SVM), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Multiple Linear Regression (MLR) to predict biomass and evaluate changes aboveground biomass over 10 years in two tropical forests in Vietnam. The results indicated that all models had good estimation performance with R2 ranging from 0.615 to 0.754. For RF, MLR, and XGBoost, vegetation indices contributed the highest model weights, occupying 77.71% – 92.48%. For ANN and SVM, textural and topographic features were the majority of the model weights (73.74 – 96.36%). The RF model performed the best using 59 variables (R2 = 0.754, MAE = 78.5 Mg·ha−1, and %RMSE = 13.57%) and ten variables (R2 = 0.745, MAE = 85.8 Mg·ha−1, and %RMSE = 16.17%). The biomass map using the RF and ten variables achieved a good degree of fitting of 0.76, so it was suitable for managing and monitoring forest biomass in Vietnam. The results indicated a sharp decrease in the areas of dense and very dense forests from 2013 to 2021 and a gradual increase in 2023.

Temporal dynamics of uncertainty and prediction error in musical improvisation across different periods

Human improvisational acts contain an innate individuality, derived from one’s experiences based on epochal and cultural backgrounds. Musical improvisation, much like spontaneous speech, reveals intricate facets of the improviser’s state of mind and emotional character. However, the specific musical components that reveal such individuality remain largely unexplored. Within the framework of human statistical learning and predictive processing, this study examined the temporal dynamics of uncertainty and surprise (prediction error) in a piece of musical improvisation. This cognitive process reconciles the raw auditory cues, such as melody and rhythm, with the musical predictive models shaped by its prior experiences. This study employed the Hierarchical Bayesian Statistical Learning (HBSL) model to analyze a corpus of 456 Jazz improvisations, spanning 1905 to 2009, from 78 distinct Jazz musicians. The results indicated distinctive temporal patterns of surprise and uncertainty, especially in pitch and pitch-rhythm sequences, revealing era-specific features from the early 20th to the 21st centuries. Conversely, rhythm sequences exhibited a consistent degree of uncertainty across eras. Further, the acoustic properties remain unchanged across different periods. These findings highlight the importance of how temporal dynamics of surprise and uncertainty in improvisational music change over periods, profoundly influencing the distinctive methodologies artists adopt for improvisation in each era. Further, it is suggested that the development of improvisational music can be attributed to the adaptive statistical learning mechanisms. This study explores the period-specific characteristics in the temporal dynamics of improvisational music, emphasizing how artists adapt their methods to resonate with the cultural and emotional contexts of their times. Such shifts in improvisational ways offer a window into understanding how artists intuitively respond and adapt their craft to resonate with the cultural zeitgeist and the emotional landscapes of their respective times.

Comparative evaluation of statistical and machine learning models for weather-driven wheat yield forecasting across different districts of Punjab

Comparative evaluation of statistical and machine learning models for weather-driven wheat yield forecasting across different districts of Punjab

Tiered maize and wheat nutrient removal coefficients estimated from available data

AbstractEstimates of cropland nutrient budgets at national to global scale generally rely on regional or global mean coefficients for quantifying nutrients removed in crop yield and by-products. Use of such mean values masks the variability in these coefficients. Using maize and wheat as examples, we assessed variation in nutrient removal coefficients, namely harvest index (HI), nitrogen (N), phosphorus (P) and potassium (K) concentrations of crop products (Grain N, Grain P and Grain K respectively) and N, P and K concentrations of crop residues (Residue N, Residue P, and Residue K respectively). Variation in these coefficients was assessed by three categories (Tiers) of estimation. Statistical (mixed-effects) and machine learning (random forest regression) models (Tier 3) were used to predict the coefficients using generally available predictor variables at a global level. Mean prediction accuracies (R2) of the mixed-effects and random forest models were 0.32 for maize coefficients and 0.45 for wheat coefficients when based on a random sub-selection of mainly replicated field experiment data. When predictions were applied to on-farm data only, prediction accuracies were lower (mean R2 values of 0.08 and 0.36 for maize and wheat respectively). Variation in, and dearth of on-farm data for the coefficients contributed to these poor prediction accuracies. Until the limitations of on-farm data are overcome, it is recommended to use Tier 2 (regional) coefficient estimates in country and global cropland nutrient balance and nutrient use efficiency estimates. Where Tier 2 values are not available, then global average (Tier 1) coefficients can be used.

Time series decomposition of land surface temperature for long-term trend forecasting and impact on nesting sea turtle habitats in the Arabian Gulf

Abstract Improving Land Surface Temperature (LST) modeling is vital for mitigating climate change effects on various ecosystems and marine habitats such as on important sea turtle habitats. Over the past decade, extreme temperatures have likely significantly affected nesting sea turtle habitats in the Arabian Gulf, with predominantly female hatchlings creating an imbalance in the sex ratio. Such shifts have profound implications for these habitats' long-term survival and conservation management. This study leverages statistical machine learning models to measure ongoing temporal variations in LST. We break down the LST time series into trend, seasonal, and noise components using classical decomposition methods like X11, SEATS, and the Seasonal and Trend decomposition using Loess (STL) approach. The long-term trends in LST data are driven by climate change rather than seasonal fluctuations. We employed Neural Network Auto Regression (NNAR), BaggedETS, Exponential Smoothing models, and STL method to project future LST values. We also explored advanced forecasting models like Dynamic Harmonic Regression, TBATS, and SARIMA for comparative performance analysis. Extended warm periods were identified for Abu Ali Island between 2017 and 2018 through several decomposition methods, likely linked to the 2015-2016 El Niño event. We also conducted a Marine Heat Wave (MHW) analysis from 2010-2020, establishing a pronounced impact of the 2015-2016 El Niño on the Arabian Gulf. In nesting beach environments with high LST, marine heatwaves could have a significant impact on sea turtle populations without human intervention such as artificially cooling the nest temperature. SARIMA model showed higher forecasting precision for in-situ weather data while NNAR model demonstrated superior performance with remotely sensed data.

Sample size effects on landslide susceptibility models: A comparative study of heuristic, statistical, machine learning, deep learning and ensemble learning models with SHAP analysis

In landslide susceptibility assessment (LSA), inventory incompleteness impacts the accuracy of different models to varying degrees. However, this area remains under-researched. This study investigated six LSA models from heuristic, statistical, machine learning and ensemble learning models (analytical hierarchy process (AHP), frequency ratio (FR), logistic regression (LR), Keras based deep learning (KBDL), XGBoost, and LightGBM) across six different sample sizes (100%, 90%, 75%, 50%, 25%, and 10%). Results revealed that XGBoost and LightGBM consistently outperformed other models across all sample sizes. The LR and KBDL models followed, while FR model was the most affected by sample size variations. AHP, an empirical model, remained unaffected by sample size. Through SHapley Additive exPlanations (SHAP) analysis, elevation, NDVI, slope, land use, and distance to roads and rivers emerged as pivotal indicators for landslide occurrences in the study area, suggesting that human activities significantly influence these events. Five time-varying indicators regarding human activity and climate validated this inference, which provides a new method to identify landslide triggering factors, especially in areas of intense human activity. Based on the findings, a comprehensive framework for LSA is proposed to assist landslide managers in making informed decisions. Future research should focus on expanding model diversity to address the effects of sample size, enhancing the adaptability of the LSA framework, deepening the analysis of human activity impacts on landslides using explainable machine learning techniques, addressing temporal inventory incompleteness in LSA, and critically evaluating model sensitivity to sample size variations across multiple disciplines.

Statistical and machine learning models for predicting spalling in CRCP

Continuously reinforced concrete pavement (CRCP), crucial for the resilience of transportation infrastructure owing to its continuous steel reinforcement, confronts a critical challenge in the form of spalling—a distress phenomenon posing a threat to pavement durability and overall structural integrity. The detachment or breakage of concrete from the surface compromises CRCP’s functionality and raises safety concerns and escalating maintenance costs. To address this pressing issue, our study investigates the multifaceted factors influencing spalling, employing a comprehensive approach that integrates statistical and machine learning techniques for predictive modeling. Descriptive statistics meticulously profile the dataset, emphasizing age, thickness, precipitation, temperature, and traffic parameters. Regression analysis unveils key relationships, emphasizing the significance of age, annual temperature, annual precipitation, maximum humidity, and the initial International Roughness Index (IRI) as influential factors. The correlation matrix heatmap guides feature selection, elucidating intricate interdependencies. Simultaneously, feature importance analysis identifies age, Average Annual Daily Traffic (AADT), and total pavement thickness as crucial contributors to spalling. In machine learning, adopting models, including Gaussian Process Regression and ensemble tree models, is grounded in their diverse capabilities and suitability for the complex task at hand. Their varying predictive accuracies underscore the importance of judicious model selection. This research advances pavement engineering practices by offering nuanced insights into factors influencing spalling in CRCP, refining our understanding of spalling influences. Consequently, the study not only opens avenues for developing improved predictive methodologies but also enhances the durability of CRCP infrastructure, addressing broader implications for informed decision-making in transportation infrastructure management. The selection of Gaussian Process Regression and ensemble tree models stems from their adaptability to capture intricate relationships within the dataset, and their comparative performance provides valuable insights into the diverse predictive capabilities of these models in the context of CRCP spalling.

Ethical considerations of use of hold-out sets in clinical prediction model management

AbstractClinical prediction models are statistical or machine learning models used to quantify the risk of a certain health outcome using patient data. These can then inform potential interventions on patients, causing an effect called performative prediction: predictions inform interventions which influence the outcome they were trying to predict, leading to a potential underestimation of risk in some patients if a model is updated on this data. One suggested resolution to this is the use of hold-out sets, in which a set of patients do not receive model derived risk scores, such that a model can be safely retrained. We present an overview of clinical and research ethics regarding potential implementation of hold-out sets for clinical prediction models in health settings. We focus on the ethical principles of beneficence, non-maleficence, autonomy and justice. We also discuss informed consent, clinical equipoise, and truth-telling. We present illustrative cases of potential hold-out set implementations and discuss statistical issues arising from different hold-out set sampling methods. We also discuss differences between hold-out sets and randomised control trials, in terms of ethics and statistical issues. Finally, we give practical recommendations for researchers interested in the use hold-out sets for clinical prediction models.

The application of chemical similarity measures in an unconventional modeling framework c-RASAR along with dimensionality reduction techniques to a representative hepatotoxicity dataset

With the exponential progress in the field of cheminformatics, the conventional modeling approaches have so far been to employ supervised and unsupervised machine learning (ML) and deep learning models, utilizing the standard molecular descriptors, which represent the structural, physicochemical, and electronic properties of a particular compound. Deviating from the conventional approach, in this investigation, we have employed the classification Read-Across Structure-Activity Relationship (c-RASAR), which involves the amalgamation of the concepts of classification-based quantitative structure-activity relationship (QSAR) and Read-Across to incorporate Read-Across-derived similarity and error-based descriptors into a statistical and machine learning modeling framework. ML models developed from these RASAR descriptors use similarity-based information from the close source neighbors of a particular query compound. We have employed different classification modeling algorithms on the selected QSAR and RASAR descriptors to develop predictive models for efficient prediction of query compounds’ hepatotoxicity. The predictivity of each of these models was evaluated on a large number of test set compounds. The best-performing model was also used to screen a true external data set. The concepts of explainable AI (XAI) coupled with Read-Across were used to interpret the contributions of the RASAR descriptors in the best c-RASAR model and to explain the chemical diversity in the dataset. The application of various unsupervised dimensionality reduction techniques like t-SNE and UMAP and the supervised ARKA framework showed the usefulness of the RASAR descriptors over the selected QSAR descriptors in their ability to group similar compounds, enhancing the modelability of the dataset and efficiently identifying activity cliffs. Furthermore, the activity cliffs were also identified from Read-Across by observing the nature of compounds constituting the nearest neighbors for a particular query compound. On comparing our simple linear c-RASAR model with the previously reported models developed using the same dataset derived from the US FDA Orange Book (https://www.accessdata.fda.gov/scripts/cder/ob/index.cfm), it was observed that our model is simple, reproducible, transferable, and highly predictive. The performance of the LDA c-RASAR model on the true external set supersedes that of the previously reported work. Therefore, the present simple LDA c-RASAR model can efficiently be used to predict the hepatotoxicity of query chemicals.