Interpretable Machine Learning Algorithm Research Articles

Algorithms for predicting COVID outcome using ready-to-use laboratorial and clinical data

The pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an emerging crisis affecting the public health system. The clinical features of COVID-19 can range from an asymptomatic state to acute respiratory syndrome and multiple organ dysfunction. Although some hematological and biochemical parameters are altered during moderate and severe COVID-19, there is still a lack of tools to combine these parameters to predict the clinical outcome of a patient with COVID-19. Thus, this study aimed at employing hematological and biochemical parameters of patients diagnosed with COVID-19 in order to build machine learning algorithms for predicting COVID mortality or survival. Patients included in the study had a diagnosis of SARS-CoV-2 infection confirmed by RT-PCR and biochemical and hematological measurements were performed in three different time points upon hospital admission. Among the parameters evaluated, the ones that stand out the most are the important features of the T1 time point (urea, lymphocytes, glucose, basophils and age), which could be possible biomarkers for the severity of COVID-19 patients. This study shows that urea is the parameter that best classifies patient severity and rises over time, making it a crucial analyte to be used in machine learning algorithms to predict patient outcome. In this study optimal and medically interpretable machine learning algorithms for outcome prediction are presented for each time point. It was found that urea is the most paramount variable for outcome prediction over all three time points. However, the order of importance of other variables changes for each time point, demonstrating the importance of a dynamic approach for an effective patient’s outcome prediction. All in all, the use of machine learning algorithms can be a defining tool for laboratory monitoring and clinical outcome prediction, which may bring benefits to public health in future pandemics with newly emerging and reemerging SARS-CoV-2 variants of concern.

A quantitative investigation on pyrolysis behaviors of metal ion-exchanged coal macerals by interpretable machine learning algorithms

Generalizing the rules from complex processes such as catalytic pyrolysis to guide their process control is always a difficult but attractive task. The influences of ion-exchange of metal ions (Na+, K+, Ca2+, Mg2+, Co2+ and Ni2+) on the pyrolysis behavior of vitrinite and inertinite from Shendong coal were investigated by thermogravimetric analyzer-Fourier transform infrared spectrometer (TG-FTIR), fixed-bed reactor (FBR), gas chromatograph-mass spectrometer (GC-MS) and X-ray diffractometer (XRD). A set of machine learning models was successfully constructed based on random forest, support vector machine, and Gaussian process regression, to quantify the relationships between pyrolysis behaviors and the properties of metals and macerals. The leave-one-out-cross validation showed that there are considerable determination coefficients (R2 > 0.9) between predicted and experimental values for most responses (17 out of 29). By combining genetic programming-based symbolic regression with the black-box algorithm, 23 symbolic regression expressions with high confidence were successfully constructed. This work is a pioneering attempt of optimization in small-scale experiments. By utilizing the highly interpretable models, a demand-orientated (multi-)optimization coal pyrolysis can be achieved. The bi-objective optimization was conducted on the yield of tar and the content of light aromatics in tar, and the results show that Co is the optimal loading metal.

Application of interpretable machine learning algorithms to predict acute kidney injury in patients with cerebral infarction in ICU

BackgroundAcute kidney injury (AKI) is not only a complication but also a serious threat to patients with cerebral infarction (CI). This study aimed to explore the application of interpretable machine learning algorithms in predicting AKI in patients with cerebral infarction. MethodsThe study included 3920 patients with CI admitted to the Intensive Care Unit and Emergency Medicine of the Central Hospital of Lishui City, Zhejiang Province. Nine machine learning techniques, including XGBoost, logistics, LightGBM, random forest (RF), AdaBoost, GaussianNB (GNB), Multi-Layer Perceptron (MLP), support vector machine (SVM), and k-nearest neighbors (KNN) classification, were used to develop a predictive model for AKI in these patients. SHapley Additive exPlanations (SHAP) analysis provided visual explanations for each patient. Finally, model effectiveness was assessed using metrics such as average precision (AP), sensitivity, specificity, accuracy, F1 score, precision-recall (PR) curve, calibration plot, and decision curve analysis (DCA). ResultsThe XGBoost model performed better in the internal validation set and the external validation set, with an AUC of 0.940 and 0.887, respectively. The five most important variables in the model were, in order, glomerular filtration rate, low-density lipoprotein, total cholesterol, hemiplegia and serum kalium. ConclusionThis study demonstrates the potential of interpretable machine learning algorithms in predicting CI patients with AKI.

Utilizing DEM and interpretable ML algorithms to examine particle size distribution's role in small-strain shear modulus of gap-graded granular mixtures

This study merges macroscale and microscale analyses with interpretable machine learning (ML) algorithm to explore the effect of particle size distribution (PSD) on the small-strain shear modulus (G0) of gap-graded granular mixtures. For this purpose, a database containing 610 groups of data samples was generated by conducting numerical triaxial compression simulations using the discrete element method (DEM). The simulations were carried out using the well-recognized DEM program PFC3D. The influence of isotropic confining pressure (p′), void ratio (e), non-plastic fines content (FC), particle size ratio (Rd), particle size ratio of coarse particles (Dmax/Dmin) and particle size ratio of fine particles (dmax/dmin) on the G0 were discussed in detail. Furthermore, the eXtreme Gradient Boosting (XGBoost) algorithm was utilized, and it was enhanced through a hyperparameter tuning process known as Bayesian optimization (BO) to improve the predictive accuracy of G0 model. Finally, the predictive outcomes of the BO-XGBoost model were interpreted using the SHapley Additive exPlanations (SHAP) method, addressing the technical gap associated with the lack of interpretability in ML models. Meanwhile, the local and global feature variable importance in nonlinear models was also elucidated to provide intuitive insight for the user. The analysis demonstrates that the outcomes from both macroscale and microscale assessments correlate well with the findings obtained through the BO-XGBoost-SHAP approach, facilitating the incorporation of ML model predictions into decision-making processes.

Application of interpretable machine learning algorithms to predict distant metastasis in ovarian clear cell carcinoma.

Ovarian clear cell carcinoma (OCCC) represents a subtype of ovarian epithelial carcinoma (OEC) known for its limited responsiveness to chemotherapy, and the onset of distant metastasis significantly impacts patient prognoses. This study aimed to identify potential risk factors contributing to the occurrence of distant metastasis in OCCC. Utilizing the Surveillance, Epidemiology, and End Results (SEER) database, we identified patients diagnosed with OCCC between 2004 and 2015. The most influential factors were selected through the application of Gaussian Naive Bayes (GNB) and Adaboost machine learning algorithms, employing a Venn test for further refinement. Subsequently, six machine learning (ML) techniques, namely XGBoost, LightGBM, Random Forest (RF), Adaptive Boosting (Adaboost), Support Vector Machine (SVM), and Multilayer Perceptron (MLP), were employed to construct predictive models for distant metastasis. Shapley Additive Interpretation (SHAP) analysis facilitated a visual interpretation for individual patient. Model validity was assessed using accuracy, sensitivity, specificity, positive predictive value, negative predictive value, F1 score, and the area under the receiver operating characteristic curve (AUC). In the realm of predicting distant metastasis, the Random Forest (RF) model outperformed the other five machine learning algorithms. The RF model demonstrated accuracy, sensitivity, specificity, positive predictive value, negative predictive value, F1 score, and AUC (95% CI) values of 0.792 (0.762-0.823), 0.904 (0.835-0.973), 0.759 (0.731-0.787), 0.221 (0.186-0.256), 0.974 (0.967-0.982), 0.353 (0.306-0.399), and 0.834 (0.696-0.967), respectively, surpassing the performance of other models. Additionally, the calibration curve's Brier Score (95%) for the RF model reached the minimum value of 0.06256 (0.05753-0.06759). SHAP analysis provided independent explanations, reaffirming the critical clinical factors associated with the risk of metastasis in OCCC patients. This study successfully established a precise predictive model for OCCC patient metastasis using machine learning techniques, offering valuable support to clinicians in making informed clinical decisions.

Does better accessibility always mean higher house prices?

Numerous studies have explored the correlations between house prices and spatial accessibility, but few have delved into the nonlinearities between both. This study uses Cardiff (UK) as a case study and applies interpretable machine learning algorithms, eXtreme Gradient Boosting (XGBoost) and SHapley Additive exPlanations (SHAP), to estimate the nonlinear effects of geometric locational accessibility and street network accessibility on house prices. The findings suggest (1) proximity to the CBD, typically the major determinant of land values in hedonic house price models, does not continuously yield higher prices; (2) street closeness centrality, a network-modelling approach to measuring accessibility, exhibits a more generalised pattern with house prices compared proximity to the CBD regardless of analytical spatial scales. The findings challenge the generalizability of Alonso’s bid-rent theory in accurately portraying the relationship between accessibility and house prices in specific urban contexts, highlighting the importance of re-evaluating classical urban theories in different city contexts using novel measures and modelling techniques.

Exploring the interrelationships between composition, rheology, and compressive strength of self-compacting concrete: An exploration of explainable boosting algorithms

This study introduces a novel methodology for enhancing the compressive strength of self-compacting concrete (SCC) via the use of the Explainable Boosting Machine (EBM), a sophisticated and interpretable machine learning algorithm. It presents a data-driven model that aims to accurately predict the strength of SCC by considering the intricate interactions among its various elements. Additionally, the model provides insights into the variables that influence SCC's compressive strength. By using EBM in conjunction with XGBoost and CatBoost algorithms, this study conducts a comparative examination of predictive abilities using datasets related to composition and rheology. The findings reveal that CatBoost has greater predictive performance using rheology dataset, as shown by an R2 value of 0.977. Conversely, XGBoost exhibits a higher predictive capability using the composition dataset, as indicated by an R2 value of 0.947. The EBM can provide comprehensive explanations at both global and local levels. It effectively identifies the key factors that have a significant influence on compressive strength. These factors include the coarse aggregate content, cement content, water content, viscosity, and V-funnel flow time. The study findings provide more evidence to support the notion that including rheological data into the model leads to a notable improvement in its accuracy. This outcome further confirms the existence of a direct correlation between rheological properties and compressive strength. The explanatory insights provided by EBM give practical instructions for customising SCC mixes to attain desired strengths. This facilitates quality control and enables personalised concrete design in the field of construction. This study highlights the potential of interpretable machine learning algorithms in improving the predictive modelling of SCC features. This advancement may lead to the development of more durable, efficient, and customised building materials.

Coevolutionary Algorithm for Building Robust Decision Trees under Minimax Regret

In recent years, there has been growing interest in developing robust machine learning (ML) models that can withstand adversarial attacks, including one of the most widely adopted, efficient, and interpretable ML algorithms—decision trees (DTs). This paper proposes a novel coevolutionary algorithm (CoEvoRDT) designed to create robust DTs capable of handling noisy high-dimensional data in adversarial contexts. Motivated by the limitations of traditional DT algorithms, we leverage adaptive coevolution to allow DTs to evolve and learn from interactions with perturbed input data. CoEvoRDT alternately evolves competing populations of DTs and perturbed features, enabling construction of DTs with desired properties. CoEvoRDT is easily adaptable to various target metrics, allowing the use of tailored robustness criteria such as minimax regret. Furthermore, CoEvoRDT has potential to improve the results of other state-of-the-art methods by incorporating their outcomes (DTs they produce) into the initial population and optimize them in the process of coevolution. Inspired by the game theory, CoEvoRDT utilizes mixed Nash equilibrium to enhance convergence. The method is tested on 20 popular datasets and shows superior performance compared to 4 state-of-the-art algorithms. It outperformed all competing methods on 13 datasets with adversarial accuracy metrics, and on all 20 considered datasets with minimax regret. Strong experimental results and flexibility in choosing the error measure make CoEvoRDT a promising approach for constructing robust DTs in real-world applications.

Translational analysis of data science and causal learning in real-world clinical evaluation of traditional Chinese medicine

Abstract Real-world clinical evaluation of traditional Chinese medicine (RWCE-TCM) is a method for comprehensively evaluating the clinical effects of TCM, with the aim of delving into the causality between TCM intervention and clinical outcomes. The study explored data science and causal learning methods to transform RWD into reliable real-world evidence, aiming to provide an innovative approach for RWCE-TCM. This study proposes a 10-step data science methodology to address the challenges posed by diverse and complex data in RWCE-TCM. The methodology involves several key steps, including data integration and warehouse building, high-dimensional feature selection, the use of interpretable statistical machine learning algorithms, complex networks, and graph network analysis, knowledge mining techniques such as natural language processing and machine learning, observational study design, and the application of artificial intelligence tools to build an intelligent engine for translational analysis. The goal is to establish a method for clinical positioning, applicable population screening, and mining the structural association of TCM characteristic therapies. In addition, the study adopts the principle of real-world research and a causal learning method for TCM clinical data. We constructed a multidimensional clinical knowledge map of “disease-syndrome-symptom-prescription-medicine” to enhance our understanding of the diagnosis and treatment laws of TCM, clarify the unique therapies, and explore information conducive to individualized treatment. The causal inference process of observational data can address confounding bias and reduce individual heterogeneity, promoting the transformation of TCM RWD into reliable clinical evidence. Intelligent data science improves efficiency and accuracy for implementing RWCE-TCM. The proposed data science methodology for TCM can handle complex data, ensure high-quality RWD acquisition and analysis, and provide in-depth insights into clinical benefits of TCM. This method supports the intelligent translation and demonstration of RWD in TCM, leads the data-driven translational analysis of causal learning, and innovates the path of RWCE-TCM.

Revisiting El-Sayed Synthesis: Bayesian Optimization for Revealing New Insights during the Growth of Gold Nanorods.

In diverse fields, machine learning (ML) has sparked transformative changes, primarily driven by the wealth of big data. However, an alternative approach seeks to mine insights from "precious data", offering the possibility to reveal missed knowledge and escape potential knowledge traps. In this context, Bayesian optimization (BO) protocols have emerged as crucial tools for optimizing the synthesis and discovery of a broad spectrum of compounds including nanoparticles. In our work, we aimed to go beyond the commonly explored experimental conditions and showcase a workflow capable of unearthing fresh insights, even in well-studied research domains. The growth of AuNRs is a nonequilibrium process that remains poorly understood despite the presence of well-established seeded growth protocols. Traditional research aimed at understanding the mechanism of AuNR growth has primarily relied on altering one reaction condition at a time. While these studies are undeniably valuable, they often fail to capture the synergies between different reaction conditions, thus constraining the depth of insights they can offer. In the present study, we exploit BO, to identify diverse experimental conditions yielding AuNRs with similar spectroscopic characteristics. Notably, we identify viable and accelerated synthesis conditions involving elevated temperatures (36-40 °C) as well as high ascorbic acid concentrations. More importantly, we note that ascorbic acid and temperature can modulate each other's undesirable influences on the growth of AuNRs. Finally, by harnessing the power of interpretable ML algorithms, complemented by our deep chemical understanding, we revisited the established hierarchical relationships among reaction conditions that impact the El-Sayed-based growth of AuNRs.

Areas simultaneously susceptible and (dis-)connected to debris flows in the Dolomites (Italy): regional-scale application of a novel data-driven approach

ABSTRACT In mountain regions, the impact of areas on the sediment conveyance can not only be described by their susceptibility to debris flow release, but also by their structural connectivity to the rivers. This generates the need to combine susceptibility and connectivity for accurate analyses of sediment transport. Our study exploits an approach developed by [Steger, er al. 2022; https://doi.org/10.1002/esp.5421] and upscales it to the South Tyrolean Dolomites region. The approach comprised the modeling of debris flow release susceptibility using an interpretable machine learning algorithm, the training of a logistic regression model, and the combination of the resultant classified maps to create a joint susceptibility-connectivity map. The results show the quantitative thresholds for the susceptibility probability and the Index of Connectivity (IC) that allow to discriminate between susceptible and not susceptible, as well as connected and disconnected areas, which are represented via a variety of maps.

Multi-omics analysis identifies potential microbial and metabolite diagnostic biomarkers of bacterial vaginosis.

Bacterial vaginosis (BV) is a common clinical manifestation of a perturbed vaginal ecology associated with adverse sexual and reproductive health outcomes if left untreated. The existing diagnostic modalities are either cumbersome or require skilled expertise, warranting alternate tests. Application of machine-learning tools to heterogeneous and high-dimensional multi-omics datasets finds promising potential in data integration and may aid biomarker discovery. The present study aimed to evaluate the potential of the microbiome and metabolome-derived biomarkers in BV diagnosis. Interpretable machine-learning algorithms were used to evaluate the utility of an integrated-omics-derived classification model. Vaginal samples obtained from reproductive-age group women with (n = 40) and without BV (n = 40) were subjected to 16S rRNA amplicon sequencing and LC-MS-based metabolomics. The vaginal microbiome and metabolome were characterized, and machine-learning analysis was performed to build a classification model using biomarkers with the highest diagnostic accuracy. Microbiome-based diagnostic model exhibited a ROC-AUC (10-fold CV) of 0.84 ± 0.21 and accuracy of 0.79 ± 0.18, and important features were Aerococcus spp., Mycoplasma hominis, Sneathia spp., Lactobacillus spp., Prevotella spp., Gardnerella spp. and Fannyhessea vaginae. The metabolome-derived model displayed superior performance with a ROC-AUC of 0.97 ± 0.07 and an accuracy of 0.92 ± 0.08. Beta-leucine, methylimidazole acetaldehyde, dimethylethanolamine, L-arginine and beta cortol were among key predictive metabolites for BV. A predictive model combining both microbial and metabolite features exhibited a high ROC-AUC of 0.97 ± 0.07 and accuracy of 0.94 ± 0.08 with diagnostic performance only slightly superior to the metabolite-based model. Application of machine-learning tools to multi-omics datasets aid biomarker discovery with high predictive performance. Metabolome-derived classification models were observed to have superior diagnostic performance in predicting BV than microbiome-based biomarkers.

Short-term prediction of celestial pole offsets with interpretable machine learning

The difference between observed and modelled precession/nutation reveals unmodelled signals commonly referred to as Celestial Pole Offsets (CPO), denoted by dX and dY. CPO are currently observed only by Very Long Baseline Interferometry (VLBI), but there is nearly 4 weeks of latency by which the data centers provide the most accurate, final CPO series. This latency problem necessitates predicting CPO for high-accuracy, real-time applications that require information regarding Earth rotation, such as spacecraft navigation. Even though the International Earth Rotation and Reference Systems Service (IERS) provides so-called rapid CPO, they are usually less accurate and therefore, may not satisfy the requirements of the mentioned applications. To enhance the quality of CPO predictions, we present a new methodology based on Neural Additive Models (NAMs), a class of interpretable machine learning algorithms. We formulate the problem based on long short-term memory neural networks and derive simple analytical relations for the quantification of prediction uncertainty and feature importance, thereby enhancing the intelligibility of predictions made by machine learning. We then focus on the short-term prediction of CPO with a forecasting horizon of 30 days. We develop an operational framework that consistently provides CPO predictions. Using the CPO series of Jet Propulsion Laboratory as the input to the algorithm, we show that NAMs predictions improve the IERS rapid products on average by 57% for dX and 25% for dY under fully operational conditions. Our predictions are both accurate and overcome the latency issue of final CPO series and thus, can be used in real-time applications.Graphical

Coupling of Solidification and Heat Transfer Simulations with Interpretable Machine Learning Algorithms to Predict Transverse Cracks in Continuous Casting of Steel

The formation of defects such as cracks in continuous casting deteriorates the quality of cast products and efficiency of steelmaking. To evaluate the risks and identify the root causes of defect formation, phenomenological quality criteria computed with a solidification and microstructure model known as InterDendritic Solidification (IDS) have previously been applied. This approach is computationally efficient and provides a fundamental perspective to defect formation in continuous casting. The aim of this work is to study the capabilities of these criteria as features in predicting transverse cracking with interpretable machine learning models. IDS is coupled with a heat transfer model known as Tempsimu to simulate the continuous casting process. Measured compositions are utilized in the simulations and defects reported at a steelmaking plant are used as labels in classification. Logistic regression, decision tree, and Gaussian Naïve Bayes classifiers are developed to predict transverse cracking in peritectic C–Mn, low‐carbon B–Ti microalloyed, and peritectic Nb‐microalloyed steels. The corresponding balanced accuracies of the best classifiers from cross‐validation are 92%, 94.6%, and 82.8%. Due to the good performance and the interpretability of the developed classifiers, the fundamental causes of transverse cracking and possibilities of avoiding it by changes in the compositions are identified.

Interpretable machine learning-assisted design of Fe-based nanocrystalline alloys with high saturation magnetic induction and low coercivity

Overcoming the trade-off between saturation magnetic induction (Bs) and coercivity (Hc) of Fe-based nanocrystalline alloys (FNAs) remains a great challenge due to the traditional design relying on trial-and-error methods, which are time-consuming and inefficient. Herein, we present an interpretable machine learning (ML) algorithm for the effective design of advanced FNAs with improved Bs and low Hc. Firstly, the FNAs datasets were established, consisting of 20 features including chemical composition, process parameters, and theoretically calculated parameters. Subsequently, a three-step feature selection was used to screen the key features that affect the Bs and Hc of FNAs. Among six different ML algorithms, extreme gradient boosting (XGBoost) performed the best in predicting Bs and Hc. We further revealed the association of key features with Bs and Hc through linear regression and SHAP analysis. The valence electron concentration without Fe, Ni, and Co elements (VEC1) and valence electron concentration (VEC) ranked as the most important features for predicting Bs and Hc, respectively. VEC1 had a positive impact on Bs when VEC1 < 0.78, while VEC had a negative effect on Hc when VEC < 7.12. Optimized designed FNAs were successfully prepared, and the prediction errors for Bs and Hc are lower than 2.3 % and 18 %, respectively, when comparing the predicted and experimental results. These results demonstrate that this ML approach is interpretable and feasible for the design of advanced FNAs with high Bs and low Hc.

A computational predictor for autophagy-related proteins using interpretable machine learning and genetic algorithm

Autophagy is a quintessential process for eliminating molecules, subcellular elements, and damaged organelles to enhance homeostasis, differentiation, development, and survival. Therefore, a thorough understanding of the sophisticated mechanism of autophagy can solely contribute to the knowledge of side effects, drug repurposing, and the development of novel poly-pharmacological strategies regarding autophagy-related diseases. Artificial intelligence approaches' broad applicability in system biology has been a promising method in identifying the autophagy-related (Art) protein, which is vitally important to regulate and control various stages of autophagy formation. Underlying explainable XGBoost and SHapley Additive exPlanations (SHAP) models, the important features and predictive model of Art protein were established. Consequently, our model performance achieved a sensitivity of 66.52 %, specificity of 82.77 %, accuracy of 77.32 %, and MCC of 0.430 via a 5-fold cross-validation evaluation. Moreover, we evaluated our model on an independent dataset, and the final results reached a sensitivity of 65.1 %, specificity of 79.1 %, and accuracy of 77.1 %. It is then observed that our model was efficient and strongly recommended for further analysis of physiological mechanisms and the development of drugs regarding autophagy. The model and dataset are freely accessible via https://github.com/khanhlee/art-predictor.

Seismic response prediction of a damped structure based on data-driven machine learning methods

Damping technology has been widely used because of its good vibration control effect. However, due to the strong nonlinearity of the added dampers, accurately predicting the seismic response of damped structures remains a challenge. This study investigates the application of interpretable machine learning (ML)-based and deep learning-based approaches to the prediction of the maximum inter-storey displacement of a damped structure. A comprehensive database consisting of 13,855 structural responses to ground motions was collected. Seven traditional interpretable ML algorithms including random forest and extreme gradient boosting (XGBoost), a convolutional neural network based on large receptive field, and seismic wave transformer (SWT) model based on a transformer network were developed. The predictions show that the error of the SWT based on unsupervised feature extraction is reduced by 50.90% compared with that of the optimal XGBoost in ensemble learning. Although the SWT has the highest global accuracy, XGBoost is found to have a smaller error when the structure is in a linear state with peak ground acceleration as partition index, so an aggregation model (AM)-based structural response prediction method was also proposed. The accuracy of the AM improved by 27.95% compared with that of the SWT. In contrast with other ML models, the proposed AM is more advantageous in terms of computational efficiency and accuracy.

A Machine Learning Approach Investigating Consumers' Familiarity with and Involvement in the Just Noticeable Color Difference and Cured Color Characterization Scale.

The aim of this study was to elucidate the relations between the visual color perception and the instrumental color of dry-cured ham, with a specific focus on determining the Just Noticeable Color Difference (JNCD). Additionally, we studied the influence of consumer involvement and familiarity on color-related associations and JNCD. Slices of ham were examined to determine their instrumental color and photos were taken. Consumers were surveyed about color scoring and matching of the pictures; they were also asked about their involvement in food, familiarity with cured ham, and sociodemographic characteristics. Consumers were clustered according to their level of involvement and the JNCD was calculated for the clusters. An interpretable machine learning algorithm was used to relate the visual appraisal to the instrumental color. A JNCD of ΔEab* = 6.2 was established, although it was lower for younger people. ΔEab* was also influenced by the involvement of consumers. The machine-learning algorithm results were better than those obtained via multiple linear regressions when consumers' psychographic characteristics were included. The most important color variables of the algorithm were L* and hab. The findings of this research underscore the impact of consumers' involvement and familiarity with dry-cured ham on their perception of color.

Predicting risk of obesity in overweight adults using interpretable machine learning algorithms.

To screen for predictive obesity factors in overweight populations using an optimal and interpretable machine learning algorithm. This cross-sectional study was conducted between June 2011 and January 2012. The participants were randomly selected using a simple random sampling technique. Seven commonly used machine learning methods were employed to construct obesity risk prediction models. A total of 5,236 Chinese participants from Ningde City, Fujian Province, Southeast China, participated in this study. The best model was selected through appropriate verification and validation and suitably explained. Subsequently, a minimal set of significant predictors was identified. The Shapley additive explanation force plot was used to illustrate the model at the individual level. Machine learning models for predicting obesity have demonstrated strong performance, with CatBoost emerging as the most effective in both model validity and net clinical benefit. Specifically, the CatBoost algorithm yielded the highest scores, registering 0.91 in the training set and an impressive 0.83 in the test set. This was further corroborated by the area under the curve (AUC) metrics, where CatBoost achieved 0.95 for the training set and 0.87 for the test set. In a rigorous five-fold cross-validation, the AUC for the CatBoost model ranged between 0.84 and 0.91, with an average AUC of ROC at 0.87 ± 0.022. Key predictors identified within these models included waist circumference, hip circumference, female gender, and systolic blood pressure. CatBoost may be the best machine learning method for prediction. Combining Shapley's additive explanation and machine learning methods can be effective in identifying disease risk factors for prevention and control.

Deciphering melting behaviors of energetic compounds using interpretable Machine learning for melt-castable applications

Melting behaviors of energetic compounds determine their thermo-responsive properties, which affects the manufacturing operations and structural stabilities. However, the decisive physicochemical mechanisms for melting behaviors of energetic compounds are absent, which hurdles the acceptancy and generalization of prediction models for rational molecular design. Herein, an interpretable machine learning (IML) approach was proposed to construct accurate prediction models for melting point values and melt-castable potentials of energetic compounds. A high-quality melting-point dataset comprising 239 neutral CHNO compounds was obtained, and the data was collected with strict criteria involving chemical formulations, crystal structures and experimental measurement operations. Tailored descriptors, which involve the priori insights for melting behaviors, were adopted for IML algorithms. The extreme gradient boosting model presents the best prediction accuracy (85 %) for the classification of melting point values between 273 K and 650 K; while 87 % accuracy is achieved for the classification of melt-castable property. Importantly, model interpretation assists to reveal that the close contacts between OC and NC, dipole, oxygen balance and the cohesive energy are decisive features for melting point prediction; while enthalpy of sublimation and polar surface area are paramount for the identification of melt-castable property. The ability to determine melting behaviors using IML will broaden the high-throughput screening of energetic compounds and offer valuable guidance towards molecular design.