High Predictive Capability Research Articles

Parameter Sensitivity Analysis for Machining Operation of Autofrettaged Cylinder Using Taguchi Method

This study investigates the impact of material parameters such as yield strength (Sy), Young’s modulus (E), and tangent modulus (T) on the safety factor (SF) of autofrettaged cylinders under 400 MPa working pressure, considering the three scenarios: no machining, internal machining, and external machining. Finite element (FE) simulations were conducted based on the Taguchi experimental design and converted into signal-to-noise (S/N) ratios to determine the optimal settings. ANOVA was utilized to evaluate the significance and percentage contributions of each factor. The analysis indicated that Sy is the most influential parameter on SF, contributing approximately 98.20% across all scenarios, including no machining, internal machining, and external machining. The contributions of E and T were minimal, but T had a slightly greater effect than E. The analytical validation of the FE model showed good agreement, with a maximum deviation of 4.37% for no machining, 4.75% for internal machining, and 5.20% for external machining. Regression analysis further confirmed the high prediction capability of the model, validated using AISI 4340 steel. The study concludes that internal machining results in higher residual stress loss compared to external machining. Overall, the analytical method tends to provide lower SF values than the numerical method, highlighting its conservative nature.

Assessing the predictive ability of information globalization under global value chains‐environmental sustainability nexus in the BRICS economies: A nonparametric causality approach

AbstractThe expansion of cross‐border information and production resources is facilitated by information globalization through the transfer of fresh ideas, products, technologies, and business models. This encourages information globalization's potential to achieve environmental and other technological advancements in the meantime and helps to make greener production possible through value‐added trade. Prior research, however, largely ignored this aspect of globalization in global value chains' studies. In order to anticipate carbon emissions (CO2) in BRICS economies, this novel study aims to assess the significance of participation in global value chains (GVCs) and information globalization (ING). The innovative research uses nonparametric causality‐in‐quantiles techniques on quarterly data from 1995Q1 to 2018Q4 to quantify for causality‐in‐variance and causality‐in‐mean because there might not be any causation at the first stage but there might be higher‐order interdependencies. The results show that GVC and ING had high predictive capability for carbon emissions, indicating asymmetry regarding environmental sustainability. Additionally, GVC and information globalization asserted that there is a significant interaction effect when it comes to forecasting pollution levels in chosen nations. The provision of financial and R&D assistance for energy efficiency and green production, as well as the use of mass and social media to raise awareness among the firms participating in global value chains, may assist in achieving SDG 13 and Cope26's goal of reducing pollution by 2030. The finding contributes crucial insights for policymakers and enhances the discourse on sustainable hones inside GVCs. The study proposes prioritizing information transparency, worldwide measures, and motivations for eco‐friendly advances to improve environmental sustainability in GVCs. Policymakers are encouraged to foster public–private associations for cohesive global endeavors in diminishing CO2 emanations inside GVCs.

Development and validation of a nomogram predictive model for cognitive impairment in cerebral small vessel disease: a comprehensive retrospective analysis.

Cerebral small vessel disease (CSVD) is a common neurodegenerative condition in the elderly, closely associated with cognitive impairment. Early identification of individuals with CSVD who are at a higher risk of developing cognitive impairment is crucial for timely intervention and improving patient outcomes. The aim of this study is to construct a predictive model utilizing LASSO regression and binary logistic regression, with the objective of precisely forecasting the risk of cognitive impairment in patients with CSVD. The study utilized LASSO regression for feature selection and logistic regression for model construction in a cohort of CSVD patients. The model's validity was assessed through calibration curves and decision curve analysis (DCA). A nomogram was developed to predict cognitive impairment, incorporating hypertension, CSVD burden, apolipoprotein A1 (ApoA1) levels, and age. The model exhibited high accuracy with AUC values of 0.866 and 0.852 for the training and validation sets, respectively. Calibration curves confirmed the model's reliability, and DCA highlighted its clinical utility. The model's sensitivity and specificity were 75.3 and 79.7% for the training set, and 76.9 and 74.0% for the validation set. This study successfully demonstrates the application of machine learning in developing a reliable predictive model for cognitive impairment in CSVD. The model's high accuracy and robust predictive capability provide a crucial tool for the early detection and intervention of cognitive impairment in patients with CSVD, potentially improving outcomes for this specific condition.

Prediction of COD in industrial wastewater treatment plant using an artificial neural network

In this investigation, the modeling of the Aksaray industrial wastewater treatment plant was performed using artificial neural networks with various architectures in the MATLAB software. The dataset utilized in this study was collected from the Aksaray wastewater treatment plant over a 9-month period through daily records. The treatment efficiency of the plants was assessed based on the output values of chemical oxygen demand (COD) output. Principal component analysis (PCA) was applied to furnish input for the Feedforward Backpropagation Artificial Neural Networks (FFBANN). The model’s performance was evaluated using the Mean Squared Error (MSE), the Mean Absolute Error (MAE) and correlation coefficient (R2) parameters. The optimal architecture for the neural network model was determined through several trial and error iterations. According to the modeling results, the ANN exhibited a high predictive capability for plant performance, with an R2 reaching up to 0.9997 when comparing the observed and predicted output variables.

Serum growth differentiation factor 15 is a novel biomarker with high predictive capability for liver cancer occurrence in patients with MASLD regardless of liver fibrosis.

Although metabolic dysfunction-associated steatotic liver disease (MASLD) patients with a Fib-4 index >1.3 are recommended for fibrosis evaluation via elastography or biopsy, a more convenient method identifying high-risk populations requiring follow-up is needed. We explored the utility of serum levels of growth differentiation factor-15 (GDF15), a cell stress-responsive cytokine related to metabolic syndrome, for stratifying the risk of clinical events in MASLD patients. Serum GDF15 levels were measured in 518 biopsy-performed MASLD patients, 216 MASLD patients for validation, and 361 health checkup recipients with MASLD. In the biopsy-MASLD cohort, multivariate analysis indicated that the serum GDF15 level was a risk factor for liver cancer, independent of the fibrosis stage or Fib-4 index. Using a GDF15 cutoff of 1.75 ng/mL based on the Youden index, high-GDF15 patients, regardless of fibrosis status, had a higher liver cancer incidence rate. While patients with a Fib-4 index <1.3 or low-GDF15 rarely developed liver cancer, high-GDF15 patients with a Fib-4 index >1.3 developed liver cancer and decompensated liver events at significantly higher rates and had poorer prognoses. In the validation cohort, high-GDF15 patients had significantly higher incidences of liver cancer and decompensated liver events and poorer prognoses than low-GDF15 patients, whether limited to high-Fib-4 patients. Among health checkup recipients with MASLD, 23.0% had a Fib-4 index >1.3, 2.7% had a Fib-4 index >1.3 and >1.75 ng/mL GDF15. Serum GDF15 is a biomarker for liver cancer with high predictive capability and is useful for identifying MASLD patients requiring regular surveillance.

Tidal modeling based on satellite altimetry observations of TOPEX/Poseidon, Jason1, Jason2, and Jason3 with high prediction capability: A case study of the Baltic Sea

This research aims to optimize the utilization of long-term sea level data from the TOPEX/Poseidon, Jason1, Jason2, and Jason3 altimetry missions for tidal modeling. We generate a time series of along-track observations and apply a developed method to produce tidal models with specific tidal constituents for each location. Our tidal modeling methodology follows an iterative process: partitioning sea surface height (SSH) observations into analysis/training and prediction/validation parts and ultimately identifying the set of tidal constituents that provide the best predictions at each time series location. The study focuses on developing 1256 time series along the altimetry tracks over the Baltic Sea, each with its own set of tidal constituents. Verification of the developed tidal models against the SSH observations within the prediction/validation part reveals mean absolute error (MAE) values ranging from 0.0334 m to 0.1349 m, with an average MAE of 0.089 m. The same validation process is conducted on the FES2014 and EOT20 global tidal models, demonstrating that our tidal model, referred to as BT23 (short for Baltic Tide 2023), outperforms both models with an average MAE improvement of 0.0417 m and 0.0346 m, respectively. In addition to providing details on the development of the time series and the tidal modeling procedure, we offer the 1256 along-track time series and their associated tidal models as supplementary materials. We encourage the satellite altimetry community to utilize these resources for further research and applications.

Precipitation data merging via machine learning: Revisiting conceptual and technical aspects

The development of accurate precipitation products of high spatio-temporal coverage is crucial for a wide range of applications. In this context, precipitation data merging (PDM), which entails the blending of satellite-based estimates with ground-based measurements, holds a prominent position, while currently there is an increasing trend in the deployment of machine learning (ML) algorithms in such endeavors. In the light of recent advances in the field, this work discusses key aspects of the PDM problem associated with: a) the conceptual formulation of the problem, that is closely related to the training of ML models and their predictive capacity, b) the selection of products fused, that is associated with the latency of final product and operational applicability of the method, c) the efficiency of single-step and two-step merging approaches, with the former one treating the problem via only regression algorithms, and the latter one via the combined use of classification and regression algorithms. By formulating PDM as a spatio-temporal prediction problem, we define and assess two different training strategies for the ML models, termed as full and per time step strategy, which entail the building of a single or several ML models, respectively. Furthermore, the performance of the full training strategy, which allows the development of predictions in both spatial and temporal dimensions, is assessed in the context of single-step and two-step merging. In each of the three scenarios, three popular ensemble tree-based ML algorithms are employed, i.e., the random forest, gradient boosting and extreme gradient boosting algorithm, resulting in nine merged products. To provide empirical evidence, we employ a datacube composed by ground-based daily precipitation observations, satellite-based and reanalysis estimates, as well as auxiliary covariates, from 1009 uniformly distributed cells (representative of a sampling area of 25 × 25 km), over four countries around the world (Australia, USA, India and Italy). The large-scale experiment indicates that: (i) full training strategy is a competitive alternative to the per time step strategy, since it enables the development of methods with improved accuracy, with respect to performance metrics and reproduction of statistics, but also with higher predictive capability and operational applicability, (ii) two-step merging enables a much better reproduction of precipitation occurrence characteristics, as reflected in the improvement of relevant categorical metrics, the reproduction of probability and autocorrelation coefficient, (iii) no significant difference was noticed in the performance of different ML algorithms.

#426 “Predict-PD”: a machine learning tool to predict patient survival in peritoneal dialysis patients

Abstract Background and Aims With the rising prevalence of chronic kidney disease (CKD), particularly among the elderly, the need for effective renal replacement therapy has become increasingly crucial. Peritoneal dialysis (PD), recognized as a gentler alternative to haemodialysis, offers advantages in cognitive preservation and overall quality of life. Patient survival on PD, a tangible outcome measure, gains particular importance, especially in an aging demographic confronting end-stage kidney disease. Predicting survival thus becomes vital in facilitating individualized treatment planning and has implications on resource allocation and patient counseling. Method Our study analysed data from the UK Renal Registry and included 22 711 incident dialysis patients choosing PD between 2007 and 2022. Our objective was to employ Artificial Intelligence (AI) algorithms in developing an advanced risk stratification indicator with high predictive capabilities, specifically tailored for use in the United Kingdom's transplant selection procedure. We conducted experiments using three distinct machine learning models and evaluated their performance in terms of calibration and discrimination, utilizing metrics such as the integrated Brier score (IBS) and Harrell's concordance index. We assessed the potential clinical utility using decision curve analysis. Results The XGBoost model demonstrated excellent predictive efficacy for patient survival, with a concordance of 0.78. Further scrutiny revealed AUC values at 1 year (0.81), 3 years (0.78), and 5 years (0.77). The integrated Brier score, a comprehensive measure of predictive accuracy, was 0.09. There was no statistical difference between the actual and predicted survival probabilities (P = 0.72). Decision curve analysis accentuated the model's clinical applicability. Conclusion This study illustrates the potential of machine learning in prognosticating patient survival in PD, offering a pathway towards individualized patient-centric management in nephrology. The findings advocate for the integration of data-driven methodologies in clinical decision-making, paving the way for enhanced personalized care in CKD management and can serve as a guide for future research.

Efficient and easily recyclable photocatalytic reduction of Se(IV) from wastewater using stable TiO2/BiOBr/cloth: Mechanism insight and machine learning modeling

Photocatalytic technology is extensively employed for the reductive removal of water contaminants; however, it contends with low catalytic efficiency and challenges in catalyst recovery. In this study, we propose integrating experimental procedures with artificial intelligence modeling to enhance the purification of Se(IV)-contaminated wastewater. We present an efficient, easily recyclable, and cost-effective strategy for photocatalyst fabrication. Specifically, we develop a novel method for Se(IV) removal by immobilizing TiO2/BiOBr onto glass fiber cloth surfaces using chitosan for Se(IV) reduction in aqueous solutions. The TiO2/BiOBr/cloth (TB-4/cloth) catalyst achieves a remarkable 99.2 % Se(IV) removal within 2 h under visible light and maintains excellent Se(IV) reduction photocatalytic activity (86.4 %) even after eight cycles, remaining easily reusable. Additionally, we develop two machine learning models, namely artificial neural network (ANN) and long short-term memory (LSTM), to validate the anticipated experimental outcomes. Both models exhibit high accuracy and predictive capability (R2 > 0.99, RMSE < 0.03). This study introduces a novel approach that combines experimentation with artificial intelligence modeling, paving the way for future advancements in Se(IV)-contaminated wastewater purification methods.

Tailored transition metal-doped TiO2@Fe3O4 nanohybrids for efficient photocatalytic dye removal: Optimization via neural networks and response surface approaches

The study focused on predicting the degradation of Remazol Turquoise Blue (RTB) dye using a photocatalyst comprised of transition metal-doped TiO2@Fe3O4. Out of the transition metal ions considered, namely Cu, Cr, Co, and Ni, Cr emerges as the most effective dopant for TiO2@Fe3O4 nanohybrid photocatalysts. This optimal performance was achieved with a catalyst loading of 50 mg and a dye concentration of 10 ppm within a 1-h reaction time, resulting in a photocatalytic efficiency of 96.70%. The optimization strategy employed a combination of Response Surface Methodology (RSM) and Artificial Neural Networks (ANN). To conduct experiments systematically, a Box-Behnken Design (BBD) rooted in RSM principles was utilized to create experimental setups for the photocatalytic degradation of the dye with the doped photocatalyst. The variables such as reaction time, catalyst dosage and dye concentration were designed to optimize the RTB dye degradation. The Regression coefficient (R2) for the developed RSM and ANN models were found to be 0.9299 and 0.9925 respectively, and revealed that the ANN has higher prediction capability and accuracy than other model. The outcomes underlined that the dosage of the catalyst prominently emerged as the principal determinant governing the efficacy of the photocatalytic activity. Remarkably, the Cr doped TiO2@Fe3O4 photocatalyst maintained its activity through six degradation cycles, with efficiencies of 99.7%, 99.4%, 99.1%, 88.8%, 87.3%, and 86.3%, respectively.

Predicting the Occurrence of Forest Fire in the Central-South Region of China

Understanding the spatial and temporal patterns of forest fires, along with the key factors influencing their occurrence, and accurately forecasting these events are crucial for effective forest management. In the Central-South region of China, forest fires pose a significant threat to the ecological system, public safety, and economic stability. This study employs Geographic Information Systems (GISs) and the LightGBM (Light Gradient Boosting Machine) model to identify the determinants of forest fire incidents and develop a predictive model for the likelihood of forest fire occurrences, in addition to proposing a zoning strategy. The purpose of the study is to enhance our understanding of forest fire dynamics in the Central-South region of China and to provide actionable insights for mitigating the risks associated with such disasters. The findings reveal the following: (i) Spatially, fire incidents exhibit significant clustering and autocorrelation, highlighting areas with heightened likelihood. (ii) The Central-South Forest Fire Likelihood Prediction Model demonstrates high accuracy, reliability, and predictive capability, with performance metrics such as accuracy, precision, recall, and F1 scores exceeding 85% and AUC values above 89%, proving its effectiveness in forecasting the likelihood of forest fires and differentiating between fire scenarios. (iii) The likelihood of forest fires in the Central-South region of China varies across regions and seasons, with increased likelihood observed from March to May in specific provinces due to various factors, including weather conditions and leaf litter accumulation. Risks of localized fires are noted from June to August and from September to November in different areas, while certain regions continue to face heightened likelihood from December to February.

A calculation method for optical properties of yolk shell based on deep learning.

The yolk shell is widely used in optoelectronic devices due to its excellent optical properties. Compared to single metal nanostructures, yolk shells have more controllable degrees of freedom, which may make experiments and simulations more complex. Using neural networks can efficiently simplify the computational process of yolk shell. In our work, the relationship between the size and the absorption efficiency of the yolk-shell structure is established using a backpropagation neural network (BPNN), significantly simplifying the calculation process while ensuring accuracy equivalent to discrete dipole scattering (DDSCAT). The absorption efficiency of the yolk shell was comprehensively described through the forward and reverse prediction processes. In forward prediction, the absorption spectrum of yolk shell is obtained through its size parameter. In reverse prediction, the size parameters of yolk shells are predicted through absorption spectra. A comparison with the traditional DDSCAT demonstrated the high precision prediction capability and fast computation of this method, with minimal memory consumption.

Prediction of Responsiveness to PD-L1/PD-1 Inhibitors Using miRNA Profiles Associated With PD-L1 Expression in Lung Adenocarcinoma and Squamous Cell Carcinoma.

MicroRNAs (miRNAs) regulate programmed cell death ligand 1 (PD-L1) and play a crucial role in tumor immune response. However, the relationship between miRNA expression patterns and PD-L1 remains unclear in both lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). We investigated PD-L1-related miRNAs that can predict treatment response in patients treated with PD-L1/PD-1 inhibitors. We selected miRNAs that were correlated with PD-L1 expression within the LUAD and LUSC datasets obtained from The Cancer Genome Atlas (TCGA) and Clinical Proteomic Tumor Analysis Consortium (CPTAC). We validated whether the miRNA profile could be used to predict the prognosis of patients treated with PD-L1/PD-1 inhibitors. Based on four public datasets, we selected 66 and 23 miRNAs associated with PD-L1 expression in LUAD and LUSC, respectively. From the above miRNAs, we identified 5 miRNAs in LUSC and 1 miRNA in LUAD that could predict the response to PD-L1/PD-1 inhibitors in a validation set of patients treated with PD-L1/PD-1 inhibitors. In LUSC, the miRNA profile exhibited a high predictive capability for the response to PD-L1/PD-1 treatment [area under the curve (AUC)=0.963] and accurately predicted prognosis (p=0.031). In LUAD, the miRNA profile was relatively less predictive than in LUSC (AUC=0.691 and p=0.213). Additionally, we observed variations in the PD-L1-associated miRNA profiles, as well as in the associated pathways, between LUAD and LUSC. The PD-L1-associated miRNA profile may predict treatment response in LUSC patients treated with PD-L1/PD-1 inhibitors and help select the PD-L1/PD-1 inhibitor treatment group.

Development of new materials for electrothermal metals using data driven and machine learning.

After adopting a combined approach of data-driven methods and machine learning, the prediction of material performance and the optimization of composition design can significantly reduce the development time of materials at a lower cost. In this research, we employed four machine learning algorithms, including linear regression, ridge regression, support vector regression, and backpropagation neural networks, to develop predictive models for the electrical performance data of titanium alloys. Our focus was on two key objectives: resistivity and the temperature coefficient of resistance (TCR). Subsequently, leveraging the results of feature selection, we conducted an analysis to discern the impact of alloying elements on these two electrical properties.The prediction results indicate that for the resistivity data prediction task, the radial basis function kernel-based support vector machine model performs the best, with a correlation coefficient above 0.995 and a percentage error within 2%, demonstrating high predictive capability. For the TCR data prediction task, the best-performing model is a backpropagation neural network with two hidden layers, also with a correlation coefficient above 0.995 and a percentage error within 3%, demonstrating good generalization ability. The feature selection results using random forest and Xgboost indicate that Al and Zr have a significant positive effect on resistivity, while Al, Zr, and V have a significant negative effect on TCR. The conclusion of the composition optimization design suggests that to achieve both high resistivity and TCR, it is recommended to set the Al content in the range of 1.5% to 2% and the Zr content in the range of 2.5% to 3%.

Assessment of chromite ore wastes for methylene blue adsorption: Isotherm, kinetic, thermodynamic studies, ANN, and statistical physics modeling

This research investigates the adsorption potential of chrysotile and lizardite, two minerals derived from chromite ore wastes, for the uptake of Methylene Blue (MB) dye from waste streams. The characterization of these minerals involves XRD, XRF, FTIR, and SEM. Results confirm the dominance of polymorphic magnesium silicate minerals, specifically chrysotile and lizardite, in the samples. The FTIR spectra reveal characteristic vibration bands confirming the presence of these minerals. The SEM analysis depicts irregular surfaces with broken and bent edges, suggesting favorable morphologies for adsorption. N2 adsorption-desorption isotherms indicate mesoporous structures with Type IV pores in both adsorbents. The Central Composite Design approach is employed to optimize MB adsorption conditions, revealing the significance of contact time, adsorbent mass, and initial MB concentration. The proposed models exhibit high significance, with F-values and low p-values indicating the importance of the studied factors. Experimental validation confirms the accuracy of the models, and the optimum conditions for MB adsorption are determined. The influence of solution acidity on MB uptake is investigated, showing a significant enhancement at higher pH values. Isothermal studies indicate Langmuir and Freundlich models as suitable descriptions for MB adsorption onto chrysotile and lizardite. The maximum adsorption capacities of MB for chrysotile and lizardite were found to be 352.97 and 254.85, respectively. Kinetic studies reveal that the pseudo-first-order model best describes the adsorption process. Thermodynamic analysis suggests an exothermic and spontaneous process. Statistical physics models further elucidate the monolayer nature of adsorption. Additionally, an artificial neural network is developed, exhibiting high predictive capability during training and testing stages. The reusability of chrysotile and lizardite is demonstrated through multiple regeneration cycles, maintaining substantial adsorption potential. Therefore, this research provides comprehensive insights into the adsorption characteristics of chrysotile and lizardite, emphasizing their potential as effiective and reusable sorbents for MB uptake from wastewater.

A preliminary study unveils CISD2 as a ferroptosis-related therapeutic target for recurrent spontaneous abortion through immunological analysis and two-sample mendelian randomization

Recurrent spontaneous abortion (RSA) affects approximately 1 % of women striving for conception, posing a significant clinical challenge. This study aimed to identify a prognostic signature in RSA and elucidate its molecular mechanisms. Prognostic gene impacts were further assessed in HTR-8/SVneo and human primary extravillous trophoblast (EVT) cells in vitro experiments. A total of 6168 differentially expressed genes (DEGs) were identified, including 3035 upregulated and 3133 downregulated genes. WGCNA pinpointed 8 significant modules and 31 ferroptosis-related DEGs in RSA. Optimal clustering classified RSA patients into three distinct subgroups, showing notable differences in immune cell composition. Six feature genes (AEBP2, CISD2, PML, RGS4, SRSF9, STK11) were identified. The diagnostic model showed high predictive capabilities (AUC: 0.966). Mendelian randomization indicated a significant association between CISD2 levels and RSA (OR: 1.069, P-value: 0.049). Furthermore, the downregulation of CISD2 promotes ferroptosis in HTR-8/SVneo and human primary EVT cells. CISD2 emerged as a pivotal gene in RSA, serving as a ferroptosis-related therapeutic target. The diagnostic model based on gene expression and Mendelian randomization provides novel insights into the pathogenesis of RSA.

Improving ventilation classification in under-actuated zones: a k-nearest neighbor and data preprocessing approach

This study investigates the use of k-nearest neighbors (k-NN) for classifying occupant positions in under-actuated zones, aiming to enhance ventilation control. The focus is on evaluating different data preprocessing techniques, particularly cumulative moving average (CMA), Kalman filtering (KF), and their combination, to boost the k-NN model's reliability and accuracy. The research uses received signal strength indicator (RSSI) data in a controlled setting. The methodology involves dividing the dataset into training and testing subsets and using root mean squared error (RMSE) to determine the best k value for model validation. The study performs a comparative analysis of the k-NN model's performance with both original and preprocessed RSSI data, focusing on metrics such as accuracy, precision, recall, F1-score, and RMSE. The findings emphasize the significant impact of the combined CMA-KF preprocessing technique in improving the model's accuracy and reliability. Specifically, this approach achieved an accuracy of 98.58%. The RMSE values are particularly noteworthy, exhibiting a perfect fit (RMSE of 0) for training data and a remarkably low RMSE of 0.119 for testing data, confirming the model's high accuracy and predictive capability.

Comprehensive study on the Python-based regression machine learning models for prediction of uniaxial compressive strength using multiple parameters in Charnockite rocks

The strength of rock under uniaxial compression, commonly known as Uniaxial Compressive Strength (UCS), plays a crucial role in various geomechanical applications such as designing foundations, mining projects, slopes in rocks, tunnel construction, and rock characterization. However, sampling and preparation can become challenging in some rocks, making it difficult to determine the UCS of the rocks directly. Therefore, indirect approaches are widely used for estimating UCS. This study presents two Machine Learning Models, Simple Linear Regression and Step-wise Regression, implemented in Python to calculate the UCS of Charnockite rocks. The models consider Ultrasonic Pulse Velocity (UPV), Schmidt Hammer Rebound Number (N), Brazilian Tensile Strength (BTS), and Point Load Index (PLI) as factors for forecasting the UCS of Charnockite samples. Three regression metrics, including Coefficient of Regression (R2), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE), were used to evaluate and compare the performance of the models. The results indicate a high predictive capability of both models. Notably, the Step-wise model achieved a testing R2 of 0.99 and a training R2 of 0.988 for predicting Charnockite strength, making it the most accurate model. The analysis of the influential factors indicates that UPV plays a significant role in predicting the UCS of Charnockite.

The Meat Standards Australia carcass grading site affects assessment of marbling and prediction of meat-eating quality in growing European beef cattle

The lack of consumer feedback on beef eating quality contributes to reduced beef consumption in Europe. The Meat Standards Australia (MSA) grading scheme can assess the palatability of beef carcasses usually graded at the 10th thoracic vertebrae. However, the European beef industry relies on late-maturing breeds usually cut at the 5th vertebrae due to commercial reasons. Data from 55 young bulls and heifers of late-maturing breeds were collected in an Italian slaughterhouse following the MSA guidelines at both carcass grading sites and sides. Intramuscular fat levels were assessed through two scores and used with other variables to feed the MSA model, which predicts the MSA index, the meat-eating quality scores (MQ4) for 5 muscles and for each carcass grading site × side combination. The scores were analyzed using a mixed linear model. A correlation analysis was conducted to predict the variables measured at the 10th site using their correspondent at the 5th carcass grading site. A stepwise regression was conducted to understand the weight of each measured variable on marbling and MQ4 scores measured both at 5th and 10th carcass grading sites. Results showed significantly higher value for the studied traits at the 5th carcass grading site, while carcass side had no significant impact. The equations had high predictive capability and MSA marbling score played a key role in explaining the variability across carcass grading sites. The differences in marbling and MQ4 scores between the carcass grading sites suggest considering this factor if the MSA grading system will be applied to Europe.

Machine learning assisted prediction of band gaps and designing of new polymers for photodetectors: A complete pipeline

The band gap is a crucial parameter of photovoltaic materials, which primarily affects their applicability and performance. Therefore, predicting band gaps of targeted materials accurately and rapidly is of utmost importance for designing photodetectors. In comparison to computationally costly first-principles calculations, machine-learned models can enable accurate, rapid, and low-cost predictions of the bandgaps. In this study, we employ machine-learning tools for the accurate and fast prediction of band gaps of polymers for photodetectors. Importantly, about fifteen (15) regression models are developed and tested to screen the models with highest predictive capability. The best models for targeted predictions among others are the light gradient boosting and hist gradient boosting models. Moreover, similarity analysis is applied to screen/search potential materials for photodetectors with high performance using reference building blocks. Gasteiger atomic charges of reference building blocks are also computed. Harvard organic photovoltaic database is explored to screen new building blocks/ monomers using their Tanimoto index. Both Breaking Retro Synthetically Interesting Chemical Substructures (BRICS) method and human controlled monomer designing methods are used to design monomers using previously searched buildings blocks automatically and manually, respectively. The band gaps of selected organic semiconductors are predicted. In addition, clustering of compounds and heatmaps of selected compounds is constructed. The synthetic accessibility score shows that BRICS-based automatic monomer designing shows negative score (i.e., difficult to synthesize), while human controlled/manual monomer designing shows positive score (i.e., easy to synthesize). Our methodology has immense potential in screening polymers with targeted bandgaps prior to experimental synthesis that could accelerate the designing of new polymers for photodetector and other photovoltaic applications.