High R2 Values Research Articles

Interpretable machine learning methods to predict the mechanical properties of ABX3 perovskites

This paper proposes the utility of interpretable ensemble learning models for predicting the mechanical properties (bulk, shear and Young moduli) of ABX3 perovskite compounds with the A, B, and X referring to the 3 elements that make the cubic 3-dimensional framework of the perovskite compounds. These models consist of 3 ensemble learning techniques namely CatBoost, Random Forest, and XGBoost. To expand the feature space, robust first-principles density functional theory calculations were used to generate some of the input features, namely elastic constants, density, volume per atom, and ground state energy per atom. The order of the input feature ranking that influences the machine learning (ML) model decisions was then determined. For this, we performed correlation analysis on the multi-dimensional input feature space, suppressed features with high collinearity, and selected features with limited correlation. We trained the three ensemble learning techniques on the desired vectorial input feature representation to predict the mechanical properties. Furthermore, we employed the Shapley Additive Explanations (SHAP) algorithm for analysing the intrinsic decision-making rationality of the ensemble learning models. We measured the performance in the context of the error metrics and coefficient of determination, R2. The results show that XGBoost outperforms other approaches when predicting the shear modulus or Young modulus of the perovskite compounds yielding the least error metrics and the highest R2 value (0.97) in the testing phase. However, both CatBoost and Random Forest outperformed XGBoost when attempting to predict the bulk modulus in the testing phase. The deficiency of the XGBoost in predicting the bulk modulus can be ascribed to an overfitting problem which can occur when the ML model gives accurate predictions for training data but not for test data. Furthermore, the SHAP algorithm provides an insight into the order of feature importance (from highest to lowest). Additionally, we conducted a post-analysis using a holistic ranking to analyse the relative importance of the SHAP feature impact comprehension for the examined ensemble learning techniques. Our findings indicate that the elastic constants are the most important input features influencing the predictive decision of the ensemble learning models.

Enhancing Solar Power Efficiency: Smart Metering and ANN-Based Production Forecasting

This paper presents a comprehensive and comparative study of solar energy forecasting in Morocco, utilizing four machine learning algorithms: Extreme Gradient Boosting (XGBoost), Gradient Boosting Machine (GBM), recurrent neural networks (RNNs), and artificial neural networks (ANNs). The study is conducted using a smart metering device designed for a photovoltaic system at an industrial site in Benguerir, Morocco. The smart metering device collects energy usage data from a submeter and transmits it to the cloud via an ESP-32 card, enhancing monitoring, efficiency, and energy utilization. Our methodology includes an analysis of solar resources, considering factors such as location, temperature, and irradiance levels, with PVSYST simulation software version 7.2, employed to evaluate system performance under varying conditions. Additionally, a data logger is developed to monitor solar panel energy production, securely storing data in the cloud while accurately measuring key parameters and transmitting them using reliable communication protocols. An intuitive web interface is also created for data visualization and analysis. The research demonstrates a holistic approach to smart metering devices for photovoltaic systems, contributing to sustainable energy utilization, smart grid development, and environmental conservation in Morocco. The performance analysis indicates that ANNs are the most effective predictive model for solar energy forecasting in similar scenarios, demonstrating the lowest RMSE and MAE values, along with the highest R2 value.

Efficacy Assessment of Molecularly Imprinted Polymer Incorporated with Starch and Macadamia Capped Silver Nanoparticles for the Removal of Multi‐Class Pharmaceuticals in Wastewater

AbstractThe presence of pharmaceuticals in water bodies has been reported to be mainly due to their incomplete removal by the conventional wastewater treatment processes, resulting to their release into the aquatic environment where they thratens to human health and aquatic life. Therefore, it is important to remove these pollutants from the aquatic environment using affordable highly efficient materials. The goal of this study was to incorporate silver nanoparticles (AgNPs) into MIPs for selective removal of sulfamethoxazole, nevirapine, and ibuprofen in wastewater. The AgNPs were synthesized using starch (St) and macadamia (MCD) as both reducing and stabilizing agents. Thermogravimetric analysis showed high stability for the St/MCD‐nanoMIPs. The ANOVA analysis revealed that the parameters affecting adsorption of St/MCD‐nanoMIPs have a significant difference (p<0.05). The St/MCD‐AgNPs experimental data fitted to both Langmuir and Freundlich adsorption isotherms, however, the non‐linear Langmuir isotherm model yielded a better fit based on the higher R2 value (0.999). The non‐linear Freundlich model best fitted for the St/MCD‐nanoMIPs. The linear pseudo‐second order model shown best fit for St/MCD‐nanoMIPs, the thermodynamic results revealed that the adsorption process was spontaneous and endothermic. The St/MCD‐nanoMIPs showed improved qualities and were highly selective (>70 %) and effective in removing the selected pharmaceuticals from wastewater.

Prediction of fresh herbage yield using data mining techniques with limited plant quality parameters

The purpose of this study was to ascertain the fresh herbage yield, fertilizer dosage, and plant characteristics of the Sorghum-Sudangrass hybrid grown in arid and semi-arid regions, as well as their interrelationships. For this reason, data from the Sorghum-Sudangrass hybrid were used to assess the predictive performance of several data mining techniques, including CHAID, CART, MARS, and Bagging MARS. Plant traits were measured in Konya and Sanliurfa during 2021 and 2022. The descriptive statistical values were calculated as follows: plant height 306.27 cm, stem diameter 9.47 mm, fresh herbage yield 10852.51 kg/da, crude protein ratio 9.66%, acid detergent fiber 33.39%, neutral detergent fiber 51.85%, acid detergent lignin 9.76%, dry matter digestibility 62.88%, dry matter intake 2.34%, and relative feed value 114.68 (average values). The predictive capacities of the fitted models were assessed using model fit statistics such as the coefficient of determination (R²), adjusted R², root mean square error (RMSE), mean absolute percentage error (MAPE), standard deviation ratio (SD ratio), and Akaike Information Criterion (AIC). With the lowest values for RMSE, MAPE, SD ratio, and AIC (246, 1.926, 0.085, and 845, respectively), and the highest R² value (0.993) and adjusted R² value (0.989), the MARS algorithm was determined to be the best model for characterizing fresh herbage yield. As a solid alternative to other data mining techniques, the MARS algorithm was shown to be the most appropriate model for forecasting fresh herbage production.

Optimization and prediction of biogas yield from pretreated Ulva Intestinalis Linnaeus applying statistical-based regression approach and machine learning algorithms

A statistical-based regression approach and machine learning (ML) algorithms (response surface methodology (RSM), feed-forward backpropagation artificial neural network (ANN) and multi-layer adaptive neuro-fuzzy inference system (ANFIS)) were explored for the optimization and prediction of biogas resulting from the anaerobic digestion (AD) of pretreated Ulva Intestinalis Linnaeus (UIL). ANFIS model was found to better predict and model the process of biogas production from the AD of pretreated UIL owing to low mean square error (MSE) and RMSE values (0.8841 and 0.9402-US, 0.9628 and 0.9812-O3, 0.1387 and 0.3724-MW and 0.3018 and 1.1410-Fe3O4) and highest values of R2 (0.9998-US, 0.9996-O3, 0.9996-MW and 0.9995-Fe3O4). Optimum conditions for biogas yield as a result of the various pretreatment processes based on the ANFIS model were US-15 power/time, time-40 min and the biogas yield-181.0 mL.gVS−1, O3-15.0 mg/min, time-40.0 min and biogas yield of 164.0 mL.gVS−1, MW-2.3 power/time, time-40.0 min and biogas yield-81.7 mL.gVS−1 and Fe3O4-20.0 mg L−1, time-40.0 min and biogas yield-154 mL.gVS−1. The results obtained show that the ML and statistical-based models were effective in approximating the biogas yield with high precision and low error and could be beneficial for the biogas production scale-up process.

Trait imputation enhances nonlinear genetic prediction for some traits.

The expansive collection of genetic and phenotypic data within biobanks offers an unprecedented opportunity for biomedical research. However, the frequent occurrence of missing phenotypes presents a significant barrier to fully leveraging this potential. In our target application, on one hand, we have only a small and complete dataset with both genotypes and phenotypes to build a genetic prediction model, commonly called a polygenic (risk) score (PGS or PRS); on the other hand, we have a large dataset of genotypes (e.g. from a biobank) without the phenotype of interest. Our goal is to leverage the large dataset of genotypes (but without the phenotype) and a separate genome-wide association studies summary dataset of the phenotype to impute the phenotypes, which are then used as an individual-level dataset, along with the small complete dataset, to build a nonlinear model as PGS. More specifically, we trained some nonlinear models to 7 imputed and observed phenotypes from the UK Biobank data. We then trained an ensemble model to integrate these models for each trait, resulting in higher R2 values in prediction than using only the small complete (observed) dataset. Additionally, for 2 of the 7 traits, we observed that the nonlinear model trained with the imputed traits had higher R2 than using the imputed traits directly as the PGS, while for the remaining 5 traits, no improvement was found. These findings demonstrate the potential of leveraging existing genetic data and accounting for nonlinear genetic relationships to improve prediction accuracy for some traits.

Comparative study of five machine learning algorithms on prediction of the height of the water-conducting fractured zone in undersea mining

Prediction of water-conducting fractured zone (WCFZ) of mine overburden is the premise for reducing or eliminating water inrush hazards in undersea mining. To obtain a more robust and precise prediction of WCFZ in undersea mining, a WCFZ prediction dataset with 122 cases of fractured zones was constructed. Five machine learning algorithms (linear regression, XGBRegressor, RandomForestRegressor, LineareSVR, and KNeighborsRegressor) were employed to develop five corresponding predictive models, taking multiple factors into account.The optimal parameters for each model are obtained through ten-fold cross-validation (10CV). The model's predictive performance was validated and assessed using two metrics, namely the coefficient of determination (R2) and mean squared error (MSE). A comparison was made with the regression performance of commonly used empirical formulas. The results indicate that the constructed model outperforms reliance solely on theoretical criteria, showing a high R2 value of up to 0.925 and a low MSE value of 3.61. The proposed model was validated in a recently established mining area on Sanshan Island, China. It shows low absolute and relative errors of 0.71 m and 2.01%, respectively, between the predicted value from the model and observation result from the field, demonstrating a high level of consistency with on-site conditions. This paves a path to leveraging machine learning algorithms for predicting the height of WCFZ.

Multi-objective optimization of an EDM process for Monel K-500 alloy using response surface methodology-multi-objective dragonfly algorithm

Monel K-500 is a high-performance superalloy composed of nickel and copper, renowned for its exceptional strength, hardness, and resistance to corrosion. To machine this material more precisely and accurately, Electrical Discharge Machining (EDM) is one of the best choices. In EDM, material removal rate (MRR) and electrode wear rate (EWR) are crucial performance parameters that are often conflicting in nature. These parameters depend on several input variables, including peak current (Ip), pulse on time (Ton), duty cycle (Tau), and servo voltage (SV). Optimizing the EDM process is essential for enhancing performance. In this research, a set of experiments were conducted using EDM on Monel K500 alloy to determine the optimal process parameters. The Box–Behnken design was used to prepare the experimental design matrix. Utilizing the experimental data, a second-order mathematical model was developed using Response Surface Methodology (RSM). R2 value is found to be 99.40% and 96.60% for MRR and EWR RSM-based prediction model, respectively. High value of R2 is indicated is indicated good adequacy for prediction. The mathematical model further used in multi-objective dragonfly algorithm (MODA): a new meta-heuristic optimization technique to solve multi-objective optimization problem of EDM. The MODA is a very useful technique to achieve optimal solutions from the multi decision criteria. Utilizing this technique, a set of non-dominated solutions was obtained. Further, the TOPSIS method was used to determine the most desirable optimal solution, which was found to be 0.0135 mm3/min for EWR and 6.968 mm3/min for MRR. These results were obtained when the optimal process parameters were selected as Ip = 6 A, Ton = 200 µs, Tau = 12, and SV = 41.6 V. Operators can machine Monel K500 by selecting the above-mentioned optimal parameters to achieve the best performance.

From organic waste to renewable energy: response surface methodology approach for optimized biodiesel production from palm weevil larvae (Rhynchophorus ferrugineus)

AbstractThis study aimed to improve the production of biodiesel from white palm larvae (WPL) and to evaluate its potential as a feedstock. The larvae were grown using kitchen and palm waste for 21 days, and the crude oil was then extracted. The extracted oil was analyzed for its physicochemical properties and fatty acid composition. The analysis found that the feedstock oil contained a high quantity of monounsaturated fatty acids, indicating its suitability for biodiesel production. To optimize the production process, response surface methodology (RSM) was employed. A central composite design was used to study four operating conditions: temperature, methanol‐to‐oil mole ratio, esterification time, and catalyst loading for esterification, as well as transesterification time and catalyst concentration. The best conditions for esterification were a methanol‐to‐oil ratio of 18:1, a temperature of 70 °C, 120 min of reaction time, and a 4% H2SO4 catalyst. For transesterification, the optimal conditions were a methanol‐to‐oil ratio of 10:1, a temperature of 65 °C, 90 min of reaction time, and a 2% KOH catalyst concentration. The experimental data fit well with models, with significant P‐values and high R2 values. The biodiesel produced under the optimized conditions met the fuel quality standards set by Brazilian, European, and American standards. In conclusion, this study successfully optimized the biodiesel production process from white palm larvae and confirmed its suitability as a feedstock. The feedstock oil had properties that were suitable for biodiesel production, and the optimized conditions resulted in biodiesel that met fuel‐quality standards.

Combining ACE, PLS-R, and SVM-R for rapid detection of adulteration in saffron samples by diffuse reflectance infrared fourier transform spectroscopy

Accurate estimation of saffron purity is essential for ensuring product integrity, particularly in the presence of common chemical adulterants such as tartrazine, sunset yellow, and quinoline yellow. This study introduces the alternating conditional expectations (ACE) algorithm, coupled with diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), as an innovative chemometric approach to address this challenge. The ACE algorithm utilizes optimal transformations for both independent and dependent variables, revealing non-linear relationships and maximizing linear effects between transformed variables. To prepare the spectral data for ACE modeling, we will perform moving average smoothing and Standard Normal Variate (SNV) transformation for preprocessing. Subsequently, we will employ the duplex algorithm to select both calibration and prediction sets from the preprocessed data matrix. The effectiveness of the ACE model was evaluated using several metrics, including the coefficient of determination (R2), adjusted R-squared (R2adj), sum of squared errors (SSE), Regression Sum of Squares (SSR) and mean squared error (MSE). ACE model exhibited excellent performance in determining saffron content even in the presence of adulterants like tartrazine, sunset yellow, and quinoline yellow, on both calibration and prediction sets. Our model achieved excellent performance on both datasets. On the calibration set, a high R2 value (0.9951) indicates a strong correlation between predicted and observed saffron content in the model. The R2adj (0.9950) further strengthens this conclusion by accounting for model complexity, suggesting the model avoids overfitting. These findings are complemented by robustness measures. The statistical parameters of the ACE model, such as total sum of squares (SST) (1.103∗104), SSR (1.098∗104), SSE (54.414), and MSE (0.878), were calculated to compare the performance of the developed methods for determining Saffron's concentration in each mixed sample.

Enhancing Stock Price Prediction in the Indonesian Market: a Concave LSTM Approach with RunReLU

This study addresses the pressing need for improved stock price prediction models in the financial markets, focusing on the Indonesian stock market. It introduces an innovative approach that utilizes the custom activation function RunReLU within a Concave Long Short Term Memory (LSTM) framework. The primary objective is to enhance prediction accuracy, ultimately assisting investors and market participants in making more informed decisions. The research methodology used historical stock price data from ten prominent companies listed on the Indonesia Stock Exchange, covering the period from July 6, 2015, to October 14, 2021. Evaluation metrics such as RMSE, MAE, MAPE, and R2 were employed to assess model performance. The results consistently favored the RunReLU-based model over the ReLU-based model, showcasing lower RMSE and MAE values, higher R2 values, and notably reduced MAPE values. These findings underscore the practical applicability of custom activation functions for financial time series data, providing valuable tools for enhancing prediction precision in the dynamic landscape of the Indonesian stock market.

Effect of Ultrasound and Osmotic Dehydration as Pretreatments on the Infrared Drying of Banana Slices.

There is a growing interest in foods with added nutritional value and extended shelf life. This study investigates the use of infrared technology in the drying of banana slices to improve their stability and quality by minimizing moisture content and water activity. The drying experiments were carried out at a temperature of 70 °C, using the following pretreatments: ultrasound-assisted (UA) immersion in water for 20 and 30 min, osmotic dehydration (OD) and ultrasound-assisted osmotic dehydration (UAOD) for 20, 30 and 40 min. The osmotic process consisted of immersing the samples in the isomaltulose solution (40.0 g/100 g deionized water) for 60 min. All mathematical models used to describe the drying process showed a good fit with high R2 values (>0.98) and low value of the relative mean error E (%), the sum of squared error and the root mean squared error. The Fick's diffusion coefficient (D eff) was higher for the samples previously treated with ultrasound for 20 and 30 min. The ultrasonic treatment resulted in shorter drying times with a reduction in average time of up to 29 %. OD was not efficient in reducing drying time, resulting in samples with lower drying rates. The samples treated with ultrasound showed less isotropic shrinkage and better color parameters. The osmotic process resulted in samples with greater rehydration capacity. The impregnation of a carbohydrate with low glycemic index in banana slices was achieved by the osmotic pretreatment, resulting in a new food product with attractive nutritional properties. This advancement represents not only a significant step in the development of functional foods, but also a major innovation in terms of processing technologies. The OD was combined with infrared drying, a method known for its superior drying rates, high heat transfer coefficient and energy efficiency. The synergy of these promising techniques not only shortens the processing times but also ensures more uniform dehydration of food products, resulting in end products that not only maintain but also optimize their nutritional value. These advances offer innovative solutions to improve food quality and also minimize environmental impact through low-energy technologies such as ultrasound and infrared treatment.

Optimizing nitrite content in powders from plasma-activated egg whites for meat preservation using response surface methodology

The study optimized the production of powders from two types of egg whites to be used as a nitrite source for meat curing, establishing a maximum utilization limit of 2%. The necessary quantity of nitrite was established through response surface methodology to guarantee its efficacy as a preservative agent and adherence to regulatory standards. High R2 values (0.9946 for hen, 0.9962 for ostrich) validated critical factors: plasma treatment time and distance, with minimal impact from drying temperature. Under specific experimental conditions (155.93 min, 16.84 cm, 41.10 °C), obtained nitrite concentration of 4000 mg kg⁻1 was confirmed through experiments. Stability tests showed that nitrite levels remained stable over four weeks under different conditions including vacuum and air-sealed packaging, as well as different temperatures (−18 °C, 4 °C, and 26 °C). Future research should investigate long-term stability, the formation of additional compounds, and the functional properties of powders to confirm their potential for food applications.

Valorization of coffee agro-industrial residue for biochar production: Use as adsorbent for methylene blue removal

This research explores using thermally modified residual coffee grounds as a bioadsorbent for removing methylene blue (MB) from aqueous solutions. The study investigates the impact of pH on adsorption capacity, emphasizing its significant influence. Kinetic studies reveal that the pseudo-first-order (PFO) model accurately describes the MB removal process, with correlation coefficients (R²) close to 1 for all samples. The Freundlich model provided the best fit for residual coffee grounds, while the Langmuir model was most suitable for biochar 1. The BiDR model provides the best fit with high R² values, indicating efficient removal for biochar 2 and biochar 3. Notably, the research shows that MB adsorption is effective without requiring pH adjustments, thereby reducing environmental impacts. The maximum adsorption capacities were 40.85, 20.35, 19.52 and 41.98 mg.g−1 using residual coffee grounds, biochar 1, biochar 2 and biochar 3, respectively. Overall, this study demonstrates the practicality of using thermally activated coffee grounds as an efficient adsorbent for removing textile dyes from effluents and highlights the potential of alternative uses for coffee grounds and biochar residues in environmental applications.

The influence of data aggregation levels on accident severity researches

Abstract This paper aims at assessing the influence of data aggregation on the estimation of the accidents’ severity. For this purpose, 27,400 accidents are investigated on three aggregation levels: individually (idl) which is highly disaggregated hourly (hrl), and daily (dil). A Severity Score (SS) is defined based on the injuries’ severity and relative cost weights. Multiple Linear Regressions are developed by assessing the SS’s variation as a collective explanatory effect of 83 predictors related to road environment, weather and lighting conditions, adherence conditions, and number of the involved vehicles. F and t tests are employed in order to assess the developed models and to investigate which variable is of statistical significance. This research shows that the higher the level of aggregation is, the higher the coefficient of determination. SS_dil model has the higher R2 value and direct relationships between SS and the number of accidents occurring in “U” curves, on roads with no shoulders, where the roadway was covered with slime or black ice and where the roadway was wet was found. One of the interesting findings of this paper shows that in the models with low R2 values (based on individual and hourly aggregated data), snow and fog significantly explain the variation of SS. As these are seasonally and short time phenomena, a further research should be focused on how the estimated models based on individual or hourly aggregated data could be improved by adding other significant predictors related to driver’s and vehicle’s characteristics and road environment’ features.

Effect of drying on bioactive compounds in Eugenia uniflora fruit pulp

AbstractEugenia uniflora is a tropical species rich in bioactive compounds that are highly sensitive to processing conditions, particularly those involving heat. Drying is a widely used method for preserving fruit pulp which can affect the stability of these bioactive compounds. The aim of this study was to assess the impact of drying at different temperatures on the retention of phenolic compounds and carotenoids present in the pulp of E. uniflora, with the goal of optimizing the drying process to preserve the pulp's quality. The E. uniflora pulp was dried in an oven with air circulation at three temperatures (45, 65, and 85°C). During the drying process, the moisture content and concentration of phenolics, β‐carotene, and lycopene were measured over time. After 260 min of drying, phenolic compounds decreased by 63.97% (45 and 65°C) and by 59.62% (85°C). Carotenoids losses were even more pronounced exceeding 89%, for all temperatures, with β‐carotene reductions of 92.91%, 90.72%, and 91.11%, at 45, 65, and 85°C, respectively. Several well‐established drying models were tested to represent the moisture content over time. Two models exhibited a high adherence to the experimental data. Zero‐order, first‐order, and second‐order degradation models were used to describe the concentrations of phenolic compounds and carotenoids. For total carotenoids, the model that showed the best results was temperature‐dependent. The first‐order model provided the best fit for β‐carotene and lycopene, with high R2 values of 0.9852 (45°C), 0.9776 (65°C), and 0.9681(85°C) for β‐carotene, and 0.9776 (45°C), 0.9715 (65°C), and 0.9659 (85°C) for lycopene. These results indicate that higher temperatures accelerate the degradation of bioactive compounds, following a predictable dynamic that can be optimized through adjustments to the drying process.Practical applicationsThe fruit of Eugenia uniflora contains phenolic and carotenoid compounds with significant nutraceutical potential. However, the production of this fruit is seasonal, necessitating effective preservation methods. Drying, which involves the removal of water from the pulp, is a common procedure aimed at extending the fruit's conservation and shelf life. Despite its benefits, the drying process poses a challenge, as bioactive compounds like phenolic and carotenoids are sensitive to thermal processing. Their degradation during drying can lead to a reduction in the fruit's bioactive potential. Therefore, it is crucial to understand the kinetics of drying and the degradation of these bioactive compounds to optimize the drying process and maximize the fruit's nutraceutical value.

The Effect of Project Management System Implementation, BIM Technology, and Cloud Collaboration on Construction Project Efficiency in Riau

This study examines the impact of Project Management System (PMS) implementation, Building Information Modeling (BIM) technology, and Cloud Collaboration on construction project efficiency in Riau, Indonesia. Utilizing a quantitative research design, data were collected from 190 professionals involved in construction projects and analyzed using Structural Equation Modeling-Partial Least Squares (SEM-PLS). The findings reveal that all three factors—PMS implementation, BIM technology, and Cloud Collaboration—positively and significantly influence construction project efficiency. Among these, BIM technology demonstrates the strongest impact, followed by Cloud Collaboration and PMS implementation. The high R² value (0.752) indicates that these factors collectively explain a substantial portion of the variance in project efficiency. This study underscores the importance of integrating advanced technological tools and management practices to enhance project outcomes in the construction industry. The results provide practical insights for construction firms in Riau and similar regions, suggesting that the adoption of BIM, Cloud Collaboration, and PMS can lead to significant improvements in project efficiency.

Grey-box solution for predicting thermo-mechanical response of rocks

Evaluating the thermo-mechanical response of rocks under high temperature treatments is crucial for various engineering geology projects. Current predictions of rock thermo-mechanical response rely on simplistic mathematical fittings treating temperature as a reduction factor, while existing machine learning algorithms often present practical challenges due to their black-box solutions. In this study, highly practical grey-box solutions, utilizing gene expression programming (GEP) are proposed for forecasting rock strength following high-temperature treatments. The dataset, comprising temperature, rock type, rock density, sample size, crack damage stress, confining pressure, and elastic modulus, serves as input parameters, with rock strength from triaxial compression tests as the output. Three grey-box solutions (mathematical formulations) based on distinct input parameter sets are proposed, all demonstrating excellent accuracy with high R2-values (R2 > 0.95) and low error values across both the training and testing phases. Feature importance analysis highlights crack damage stress, confining pressure, and elastic modulus as statistically significant parameters influencing the strength of rocks subjected to high temperatures. External validation of the proposed models indicates strong generalization capabilities, underscoring their ability to perform well beyond the training data. Furthermore, a monotonicity study demonstrates that the proposed models align with the expected physical processes. The proposed formulations offer valuable field implications, effectively addressing the limitations of labor-intensive and costly laboratory processes for evaluating rock thermo-mechanical responses.

REMOVAL OF VARIOUS METAL IONS IN WATER BY DIFFERENT PRE-TREATMENTS OF FLY ASH

Rapid urbanisation in Malaysia has accelerated water pollution in rivers and other water sources, causing irreversible harm to the ecosystem. In view of that, this study aimed to work on using fly ash to address certain heavy metal components (chromium (Cr), copper (Cu), nickel (Ni), and zinc (Zn)) present in polluted water. The experiment employed three batches of fly ash. Two batches were treated with sodium hydroxide (NaOH-FA) and hydrochloric acid (HCl-FA), whereas one batch was left untreated (UFA). The three batches of adsorbents were examined by using a jar test after solutions containing 100 mg/L of Cr, Cu, Ni, and Zn ions were made. The results of various contact periods demonstrated that the fly ash had variable capacities for metal ion adsorption. The maximum adsorption of UFA was 79.958%(Cr), 80.814%(Cu), 81.580%(Ni), and 82.742%(Zn) while HCl-FA was adsorbing 77.148%(Cr), 82.546%(Cu), 78.896%(Ni), and 78.248%(Zn). NaOH-FA in this study was found to adsorb 80.828%(Cr), 79.230%(Cu), 81.692%(Ni), and 77.394%(Zn). Further to this, it was revealed that the Temkin Isotherm model was best fitted with the highest R² values (> 0.98). The negative value of the slope, B indicated that the adsorption is an endothermic process which leans towards physical adsorption. In conclusion, this study demonstrated the successful application of fly ash in water or wastewater treatment of metal ions.

Parametric modeling of resin-bonded sand mold systems using machine learning-based approaches

This study presents the experimental investigations on modeling of the compression strength and permeability of the resin-bonded sand mold system through Machine Learning approaches. The process of constructing the Fuzzy Logic system was automated by utilizing a data-base and rule-base optimized through genetic algorithms, and the recorded datasets. This research employed three AI models, namely Artificial Neural Networks (ANN), Decision Tree (DT), and Random Forests (RF), using the datasets produced by the GA-tuned Fuzzy model. The objective of this study was to assess and compare the predictive capabilities of three different AI models (Artificial Neural Networks, Decision Tree, and Random Forests) in terms of predicting the values of Compression Strength and Permeability. The complete dataset was divided into two separate subsets, referred to as training data and testing data. Based on the findings, it appears that Random Forest (RF) model exhibits promising potential in accurately predicting the desired mold qualities. The model achieved a high R2 value of 0.9487, indicating a strong correlation with the target values. Additionally, the model demonstrated impressively low Mean Squared Error (MSE) and Mean Absolute Error (MAE) values of 117 and 17.6 points, respectively. Expanding the dataset size may further enhance the efficacy of the models.