Coefficient Of Determination R2 Research Articles

Derivation of Analytical Equations for the Fundamental Period of Framed Structures Using Machine Learning and SHAP Values

The fundamental period is one of the most important parameters for the design of new structures as well as for estimating the capacity of existing ones. Thus, to estimate it, various design codes and researchers have adopted several approximate analytical equations based on a number of key structural parameters. To this end, the present study introduces a novel methodology for deriving the analytical equations for the fundamental period of reinforced concrete structures. The methodology is based on machine learning explainability techniques, specifically the so-called SHapley Additive exPlanations values. These values are commonly employed as an explainability tool. However, in the proposed novel approach they are employed as a basis to fit analytical curves, which allows the resulting equations to be constructed sequentially and in an informed manner while controlling the balance between accuracy and complexity. An extended dataset consisting of 4026 data points is employed, on which a Gradient Boosting Machine model is fitted. The model achieves excellent accuracy, with a coefficient of determination R2≈0.99, while the equations derived from the proposed formulation achieve an R2≈0.95 and Mean Absolute Error ≈0.12. This demonstrates the potential applicability of the proposed methodology in a wide array of similar engineering challenges.

Model for Estimating Soil Chemical Properties with RGB Drone Images

Precision agriculture optimizes crop management by providing accurate data on soil chemical properties, thereby improving agricultural productivity and sustainability. This study aims to develop models to estimate soil chemical properties, such as pH, electrical conductivity (EC), and organic matter (OM), by analyzing drone-captured RGB images. The methodology included photogrammetric flights with a DJI Phantom 4 Pro drone equipped with a 20 Mpx camera and simultaneous sampling, laboratory analysis and on-site measurements, with Royal Eijkelkamp EC meter set voor grond multiparameter sensors and pH meter set for soil and water. The aerial images were processed with the PIX4Dmapper software, to generate the orthophoto and spectral bands. With the resulting orthophoto of 1.6 cm/pixel, eight spectral indices were calculated, using the spatial analysis tools of ArcGIS software. The in situ results showed an average pH value of 5.83, indicating a slightly acidic soil, and an EC of 1.09 dS/m, suggesting a soil with a low concentration of dissolved salts. Laboratory analyses showed a medium-high content of OM, with an average of 5.19 %. A strong correlation was found between OM and pH_index with coefficients of determination R2=0.55, while moderate correlations were also observed between pH with pH_index and EC with sal_index6 with coefficients of determination R2=-0.39 and R2=0.42 respectively. The aforementioned results allowed the generation of two models for the estimation of these variables from RGB images.

Characteristics of gusts with different velocity profiles and control parameters

The characteristics of gust flow are essential for gust response and alleviation. To investigate the influence of control parameters on gusts with different velocity profiles, four vertical gust profiles were designed. Methods were proposed to generate them with two pitching airfoils in a low-speed water tunnel. The velocity field was measured via phase-locked particle image velocimetry. The coefficient of determination R2 was proposed to evaluate the generated gust profile quality, which referred to the quality of the vertical velocity profile. The influence of control parameters on different gust profiles was investigated, and the cause of the profile distortion was explored. For continuous sine gusts, the gust ratio GR increased approximately linearly with the pitching amplitude, while the gust ratio initially increased and then decreased with increasing frequency. As the two control parameters increased, the flow uniformity decreased because the airfoil wakes disturbed the measured flow field. In terms of continuous 1-cosine gusts, the gust ratio increased nonlinearly with pitching amplitude. Compared with those of the sine gusts, the GR values of the 1-cosine gusts were higher, whereas the R2 values were lower. In addition, the discrete and continuous gust profiles had similar distortion near the peaks. However, discrete gusts had lower R2 values than continuous gusts because the starting and stopping vortices of the pitching airfoils disturbed the gust flow. Based on these findings, a method to improve the profile quality and field uniformity by increasing the spacing of the pitching airfoils was proposed. This work can support further studies of gust response and alleviation during complex gust encounters.

IoT and AI: a panacea for climate change-resilient smart agriculture

The application of Internet of Things (IoT) and Artificial Intelligence (AI) for disaster preparedness and sustainable agriculture has been a topic of great interest lately. In the last few years, extreme weather swings due to climate change caused by global warming have caught the farming community off guard, especially in the developing world. One of the key objectives of smart agriculture is optimal use of freshwater, which has become an increasingly scarce resource around the world. Reference Evapotranspiration (ETo), an estimation of total flux of water evaporating from a reference surface is an important parameter for irrigation management. IoT & AI-based location-specific estimation of ETo for crop water requirements augments the decision-making process. In this work, we utilize the Hargeaves and Samani (H–S) model and six regression algorithms for the estimation of ETo. We create a location-specific dataset with locally sensed IoT data from a flood warning system and remotely sensed meteorological data, spanning over 5 years. We train and test Linear Regression (LR), Multilayer Perceptron (MLP), Radial Basis Function (RBF), Support Vector Regression (SVR), Bagging and Random Forest (RF) algorithms on the locally curated dataset with 20 basic, extracted, and derived attributes. We gradually reduce number of attributes in the dataset from 20 to 3 and compare performance of the six algorithms using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Relative Absolute Error (RAE), Root Relative Squared Error (RRSE), Coefficient of Determination R2, Kendall Tau and Spearman Rho metrics. SVR shows superior performance with an MAE of 0.03 and an RMSE of 0.05, followed closely by MLP with an MAE of 0.04 and RMSE of 0.06 with a dataset of 12 attributes. The performance of Bagging and RF algorithms remains relatively unchanged with feature reduction whereas RBF shows slight improvement in performance when number of attributes is reduced to 3. Finally, we develop a novel ensemble hybrid model using the Stacked Generalization technique, which outperforms all individual models in prediction accuracy when using reduced-feature datasets. This work clearly delineates the performances of a diverse set of ML algorithms for feature-rich and feature-scarce scenarios and demonstrates the efficacy of our hybrid ensemble ML algorithm for estimating ETo under limited availability of data in resource-constrained environments.

Empirical Correlation of Hossaena Town Soil Index Properties with California Bearing Ratio (CBR) Values

California bearing ratio (CBR) is a key input factor in pavement structural design that is used to determine character size, the strength of unbound materials, and subgrade soils. The construction of a straightforward and logical regression model to forecast the CBR of unbound materials as a function of the index properties of the material. This work explains the relationship between compaction parameters and Atterberg limit tests. Laboratory test was conducted to get the independent (percent finer, liquid limit, plastic limit, plastic index, maximum dry density and moisture content), and dependent variables (CBR). A statistical package for social science software (SPSS) used to develop the mathematical model both in single and multiple regression. The viability of using multiple linear regression analysis to relate CBR values to soil index features was investigated. In this study, SPSS is used to examine each independent variable's significance individually. The multiple linear regression analysis was used to create the correlation, which had an eighteen-person sample size and a modest determination coefficient of R2 = 0.821. The CBR has good relation with maximum dry density and optimum moisture content compared to other parameters. This model applicable for small structure in the specified soil type, MH.

Assistance of aminated straw on enhanced biodegradation of phenol through bacteria and potential prediction of machine learning for bioremediation

ABSTRACT How to treat phenol-containing wastewater harmlessly is an urgent problem to be solved. In this study, a kind of biomaterial was prepared through strain PAO1 immobilized on aminated straw to enhance the phenol removal rate and efficiency. The aminated straw-assisted PAO1 to increase the phenol degraded concentration from 1900 to 1,500 mg/L, and shorten the degraded time by 44 h at 1,500 mg/L phenol. The immobilized PAO1 could remove phenol at pH 10 and 11, which was 2.7 and 3.8 times higher than free bacteria, respectively. In addition, the immobilized PAO1 could totally remove phenol, which was twice as high as that of free bacteria, at 4% NaCl stress. Furthermore, the removal efficiency of immobilized PAO1 was higher than that of free bacteria under the stress of various metal ions, especially for Co2+ and Pb2+. The determination coefficients R2 and root mean square error showed that the back propagation artificial neural network model could predict the degradation of phenol under various conditions, saving time and economic cost. The present study can be envisioned that this biomaterial has great potential in the bioremediation of organic pollutions.

Melt-blown polypropylene nonwoven as an efficient and eco-economic sorbent for pipette tip micro-solid phase extraction for the determination of tyrosine kinase inhibitors

BackgroundThe detection of tyrosine kinase inhibitors (TKIs) in biological fluids is essential due to their critical role in cancer therapy and the variability in individual drug metabolism, which necessitates precise dosing. Traditional methods for analyzing TKIs in biological fluids, such as blood plasma, typically involve complex sample preparation techniques that can be resource-intensive, environmentally burdensome, and not sufficiently sensitive for low-concentration analytes. There is a pressing need for more efficient, economical, and environmentally friendly methods that can enhance sensitivity and throughput without compromising accuracy. ResultsThis study explores the use of melt-blown polypropylene nonwoven (MBPP), commonly found in face masks, as a novel sorbent for pipette-tip micro-solid phase extraction (PT-μSPE). MBPP demonstrated excellent hydrophobicity and significant mesoporous adsorption capacity. An extraction device was fashioned by inserting a segment of MBPP (15 mg) into a 200 μL disposable plastic pipette tip, which was then attached to a 2.5 mL disposable plastic syringe. The MBPP's fabric form removes the need for a frit, allowing the extraction process to be completed in just 3 min through simple plunger manipulation. The method achieved extraction recoveries ranging from 60.5% to nearly 100%. Subsequent method validation using liquid chromatography-tandem mass spectrometry (LC-MS/MS) showed satisfactory linearity (coefficient of determination R2 > 0.993), accuracy (relative recoveries: 86.3%–114.8%), and precision (relative standard deviation: 3.4%–11.3%), with detection limits between 0.022 and 0.135 ng mL-1. SignificanceThe introduction of MBPP for PT-μSPE represents a significant advancement in the bioanalytical detection of TKIs, offering a highly efficient, cost-effective, and environmentally sustainable method. It compares favorably with existing techniques, offering advantages in terms of cost, environmental impact, and ease of use. This approach has the potential to be widely adopted for routine monitoring of TKIs in clinical settings.

Influence of degree of saturation on the thermal conductivity of soils: Experimental and comparative model analysis

The thermal conductivity can characterize thermal behaviour of soils in different engineering studies: environmental, geothermal, geotechnical and buildings construction. However, accurately measuring and predicting this parameter poses a challenging task. Measuring this parameter can be very complex, considering several factors, such as soil heterogeneity and external climatic conditions. In addition, predictions models may not capture all the nuances of real-world soil conditions, leading to less accurate predictions of ?. The objectives of this paper encompassed two primary aspects: (i) Evaluation of laboratory tests of thermal conductivity of unsaturated soil. (ii) Assessment of five highly recommended soil thermal conductivity models to determine their validity and strengthen their trustworthiness. The studied material consisted of calcareous tufa locally available in Beni-Saf region (Algeria). The tested samples were compacted to the Standard and Modified Proctor Optimum (SPO, MPO) followed by drying periods under laboratory conditions. The thermal conductivity of samples was evaluated using transient method. The models’ predictive results of the thermal conductivity were assessed with different criteria such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Coefficient of Determination R². Furthermore, this work also attempts to deliver an in-depth discussion of the effect of degree of saturation Sr on the thermal conductivity of soils.

The Effect of Biofertilizer Type and Chicken Manure Dosage on the Growth and Yield of Kale (Brassica oleraceae L.)

Aims: To study the effect of the type of biofertilizer and the dose of chicken manure on the growth and yield of kailan (Brassica oleraceae L.). Study Design: Experimental design, using factorial randomized block design consisting of two factors with three replications. Place and Duration of Study: Experimental Station of the Faculty of Agriculture, Universitas HKBP Nommensen in Medan, from June to October 2023. Methodology: The research used a factorial randomized block design consisting of two factors, namely the type of biofertilizer (P) and the dose of chicken manure (A), with three replications. The type of biofertilizer consisted of four levels, namely: P0 = No application of biofertilizer (control), P1 = IMO of pineapple peel (100 ml/l water), P2 = Eco-enzyme (1 ml/liter water), and P3 = Bio-Extreme biofertilizer (5 ml/m2). The dose of chicken manure (A) consisted of three levels, namely: A0 = 0 tons/ha (control), A1 = 15 tons/ha, and A2 = 30 tons/ha. Biofertilizer according to the treatment level was applied four times at 1 week before transplanting and 1, 2 and 3 weeks after transplanting, while the chicken manure was applied at 1 week before transplanting. The parameters observed were: plant height, number of leaves, root volume, root length, wet weight per plot, selling weight per plot, and yield per hectare. Results: Both the type of biofertilizer and the dose of chicken manure treatments had no significant effect on all parameters, whereas the interaction between the type of biofertilizer and the dose of chicken manure had a significant effect on root length, but had no significant effect on other parameters. The relation between root length and the dose of chicken manure using IMO of pineapple peel was positive linear, with mathematical equation yP1 = 0.1733 x + 6.3833, with coefficient of determination R2 = 0.0788 and coefficient of correlation r = 0.2807, while the one using EE was positive quadratic, with mathematical equation yP2 = -0.3867 x2 + 0.5733 x + 7.25, with coefficient of determination R2 = 1 and coefficient of correlation r = 1. Conclusion: The interaction between the type of biofertilizer and the dose of chicken manure had a significant effect on the root length of kale but had no significant effect on plant height and number of leaves at 4 WAT, root volume, wet weight per plot, selling weight per plot and yield per hectare. The best combination was obtained in P2A1 (eco enzyme as biofertilizer and chicken manure dose of 15 tons/ha) but it was not significantly different from the combinations of P1A0, P1A1, P2A2, P3A0, P3A1 and P3A2. With eco enzyme treatment, the relationship between the dose of chicken manure and root length was positive quadratic with the optimum dose of chicken manure for a maximum root length of 7.5 cm being 7.4 tons/ha. The dose of chicken manure had a significant effect on root length, but had no significant effect on other parameters, while the type of biofertilizer had no significant effect on all parameters.

Comparison of machine learning algorithms on a low heat rejection diesel engine running on ternary blends

Depletion of fossil fuels and increasing the energy demand are the critical concerns for the sustainable growth of every country. These challenges highlight the need for a cleaner fuel globally. In particular, substituting the petroleum fuels with biofuels could significantly support sustainability. Hence, this study explores the impact of acetylene induction in a low heat rejection (LHR) diesel engine with ternary blend (TB) as a pilot fuel. TB includes 10%methanol + 20%WCOB + 70%diesel. During the dual fuel operation, the acetylene induction was varied (12, 18, and 24 lpm) with different pilot fuel combinations in a LHR engine. From the experimental results, it was revealed that brake thermal efficiency (BTE) is improved by 10.3% for TB with acetylene at 24 lpm and increased the exhaust gas temperature (EGT) by 10.3% at full load conditions. This study further evaluates the implementation of machine learning algorithms, namely, Random Forest regression (RFR) and Polynomial regression (PR) in predicting the performance parameters [BTE, brake-specific fuel consumption (BSFC), and EGT]. The results indicate that RFR outperforms PR in accurately predicting engine performance characteristics with a coefficient of determination R2 = 0.97, R2 = 0.98 for BTE, whereas for BSFC the R2 = 0.96, R2 = 0.95, and for EGT, the R2 = 0.95, R2 = 0.97 for the test fuels diesel and TBA3, respectively. These findings suggest that the combination of LHR technology, TB fuel, and acetylene can enhance engine efficiency and performance characteristics, and machine learning models can effectively predict these outcomes.

The Influence of Work Discipline, Organizational Culture, Quality of Work Life, and Work Motivation on Job Satisfaction (Study of SL Corp Indonesia Employees in the Purwokerto Area)

This research aims to determine the influence of work discipline, organizational culture, quality of work life, and work motivation on job satisfaction among SL Corp employees. Using Quantitative Methods with a population of 154 employees and a sample of 111 employees. Questionnaires were distributed in March-May. This research includes descriptive statistics, reliability, and basic assumptions: normality, multicollinearity, and heteroscedasticity. Model feasibility, multiple linear regression test, partial, and coefficient of determination R2 using the SPSS 26 analysis tool. The results of the analysis show that work discipline, organizational culture, and quality of work life have no effect on job satisfaction. Meanwhile, work motivation has a significant positive effect on job satisfaction and has important implications in managing human resources and increasing satisfaction among SL Corp employees.

Artificial Neural Network Model to Predict the Factor of Safety in Earth Dams Subjected to Rapid Drawdown

Rapid drawdown has been identified as one of the most frequent causes of slope failures due to the effects associated with drought and operational changes when incorporating hydroelectric plants, which influence the filling level of earth dams. The main goal of this research is to obtain predictive models based on Artificial Neural Networks that return the factor of safety of the upstream slope in homogeneous earth dams in the face of the effect of rapid drawdown. Three geometries and 40 soils were defined to form the embankment, from which hybrid numerical models of transient water flow with unsaturated soils were built, considering three discharge speeds. From these results, a database was built to develop the predictive models, by means of the KNIME program and an algorithm based on Artificial Neural Networks. The behavior of the factor of safety as a function of time is also analyzed to establish its recovery intervals. Main results show that the minimum factor of safety is obtained between 52 % and 88 % of the total drawdown time. Regarding the predictive models, the adjusted R2 determination coefficients were greater than 95 % in all cases and the errors remained below 10 %. This demonstrates a high effectiveness of this algorithm and the validity of its application to geotechnical problems.

ANALYSIS OF STAFFING OF THE MEDICAL AND GENETIC SERVICE IN UKRAINE

Aim. To study the staffing conditions of the medical-genetic service in Ukraine during the period from 2008 to 2022 and to forecast the staffing dynamics until 2026. Materials and Methods: Statistical data from healthcare institutions in Ukraine that provided medical-genetic services during the period from 2008 to 2022 were analyzed, based on reporting forms 21, 17, 49 from the City Health Center website http://medstat.gov.ua/ukr/statdan.html and http://medstat.gov.ua/ukr/statdanMMXIX.html. Single-factor linear regression models were applied for mathematical modeling, with quality assessment conducted using the coefficient of determination R². Calculations were performed using MedCalc® Statistical Software (v.22.009). Results: Compared to 2008, there was a significant reduction in the number of full-time positions of heads of offices (genetic doctors), genetic doctors, and genetic laboratory doctors in Ukraine in 2022 by -45.24%, -24.83%, and -23.39%, respectively. A significant decrease in the number of these positions was determined (p<0.05). In 2022, compared to 2008, there was also a decrease in the staffing rate of all full-time positions with occupied positions of the personnel of centers/offices. Specifically, for genetic doctors, the staffing rate in 2022 was 69.62% compared to 85.37% in 2008, and for genetic laboratory doctors, it was 57.54% compared to 94.62% in 2022 and 2008, respectively. The level of provision of geneticists (individuals) in Ukraine in 2022 was 1.99 per 1 million population, which is -16.03% less than in 2008. Furthermore, compared to 2008, the share of vacant positions for genetic doctors in 2022 increased 2-fold, and for genetic laboratory doctors almost 8-fold. According to the results of mathematical forecasting, in 2026, a further decrease in the availability of genetic doctors to the level of 1.77 per 1 million population is predicted (R²=0.88, p<0.001). Results. Compared to 2008, there was a significant reduction in the number of full-time positions of heads of offices (genetic doctors), genetic doctors, and genetic laboratory doctors in Ukraine in 2022 by -45.24%, -24.83%, and -23.39%, respectively. A significant decrease in the number of these positions was determined (p<0.05). In 2022, compared to 2008, there was also a decrease in the staffing rate of all full-time positions with occupied positions of the personnel of centers/offices. Specifically, for genetic doctors, the staffing rate in 2022 was 69.62% compared to 85.37% in 2008, and for genetic laboratory doctors, it was 57.54% compared to 94.62% in 2022 and 2008, respectively. Furthermore, in comparison with 2008, the share of vacant positions for genetic doctors in 2022 increased in 2-fold, and for genetic laboratory doctors almost in 8-fold. According to the results of mathematical forecasting, in 2026, a further decrease in the availability of genetic doctors to the level of 1.77 per 1 million population is predicted (R²=0.88, p<0.001). Conclusions. Based on the analysis of staffing of the medical-genetic service in Ukraine during the period 2008-2022, conclusions can be drawn about the reduction in the number of full-time positions and the availability of genetic doctors in the medical-genetic service in Ukraine. To ensure proper medical-genetic counseling for the population, it is necessary to develop an effective strategy for improving the organization of the medical-genetic service, capable of providing comprehensive support for patients with genetic pathologies.

Utilizing erythrosine B absorption spectrum shifts for quantitative determination of octreotide and bromocriptine in their pure forms and pharmaceutical formulations. Evaluation of the method greenness

In the present study, two drugs for treating acromegaly are investigated which are either synthetic growth hormone inhibiting hormone (GHIH) or dopamine receptor agonist. Octreotide (OCT) and bromocriptine (BCT) were quantified with a quick, simple, and sensitive spectrophotometric approach in their authentic forms and commercial dosage forms. This approach was based upon formation of ion pair complex between erythrosine B and investigated drugs in an aqueous buffered solution. For OCT determination the higher sensitivity was obtained using ΔA at 525 nm. On the other hand, BCT was estimated using the absorbance at 556 nm. Under optimal conditions for the reaction, construction of calibration curves were performed within concentration range of 0.4–4 µg ml−1 for OCT and 1–9 µg ml−1 for BCT with with low detection limits (0.094 and 0.235 µg ml−1) and low quantitation limits (0.29 and 0.71 µg ml−1) for OCT and BCT respectively. The developed approach was assessed for linearity, accuracy, precision, limits of detection, and limits of quantitation in compliance with ICH validation recommendations. The suggested approaches demonstrated good linearity, as evidenced by determination coefficients (r2) of 0.999 with a slope 0.14 for OCT and 0.9989 with a slope 0.06 for BCT. Additionally, the environmental friendliness of the investigated method was assessed with some of the recent green analytical chemistry metrics.

Influence of the Addition of Palm Kernel Shell and Ash on Concrete Performances: Study of Correlations between Intrinsic Material Properties

Sustainable development objectives in the construction sector is currently hampered by the high cost of materials (cement, aggregates) and the environmental impact of waste in some developing countries. The obtained results in this work on the influence of ash (PKSA) and palm kernel shell (PKS) used as a partial addition of cement and aggregates in concrete composition, enable their valorization as local materials to manufacturing the lightweight concrete with low-cost. This is an interesting contribution to the development of sustainable construction and environmental protection. The used palm kernel shells are produced in the palm oil industry in Republic of Congo. The highest values for density Cd(2348kg/m3), compressive strength Cs(27MPa) and splitting tensile strength Ts(2.4MPa) for concrete using PKSA were obtained at 2.5%. Those for concrete using PKS were obtained at 5 %, i.e. Cd(2165kg/m3), Cs(22MPa) and Ts(1.90MPa). the increase in concrete properties with PKSA compared to PKS is explained by the pozzolanic reaction of the palm kernel ash, which acts as a hydraulic binder. Correlations between fundamental concrete properties reliably describe the influence of PKS and PKSA, with a coefficient of determination R2 (0.9) superior at 0.5; the obtained mathematical models to prediction the concrete properties are a significant contribution for engineers. PKS is a concrete plasticizer, PKSA is a concrete setting and hardening accelerator. The production of low-carbon concrete with PKSA addition is a major step forward for the concrete industry.

Development of HPTLC method for simultaneous determination of quercetin and kaempferol in leaf extract of Hibiscus mutabilis

The aim of this study was to develop and validate a densitometric High-Performance Thin-Layer Chromatography (HPTLC) method for the simultaneous quantification of quercetin (Q) and kaempferol (K) in Hibiscus mutabilis leaf extracts. The analyses were performed on silica gel 60 F254 plates using a mobile phase composed of toluene, formic acid, and ethyl acetate (6:0.4:4, v/v/v). Detection was carried out at a wavelength of 272 nm using a deuterium and tungsten light source. The method exhibited excellent linearity over the concentration range of 100–600 ng/spot for quercetin and 500–3000 ng/spot for kaempferol, with determination coefficients (r2) of 0.9989 and 0.9973, respectively. The method showed no interferences from the plant matrix. The relative standard deviation (RSD) values for intra- and inter-day precision were less than 2% for both flavonoids. Recovery rates ranged from 97.69% to 99.20% for quercetin and from 89.91% to 95.87% for kaempferol. The limits of detection (LOD) were 190.23 ng/spot for quercetin and 187.23 ng/spot for kaempferol, while the limits of quantification (LOQ) were 570.10 ng/spot for quercetin and 566.12 ng/spot for kaempferol. This validated HPTLC method is reliable, precise, and accurate, making it suitable for the quality control of Hibiscus mutabilis leaf extracts. The study’s findings can be broadly applied to the quality control of herbal products, pharmacological research, and the development of nutraceuticals. The method’s ability to provide rapid and accurate quantification makes it an invaluable tool for researchers across various disciplines.

The Influence of E-Service Quality and E-WOM on Customer Satisfaction of Shopee Users (Study on Business Administration Students at Bali State Polytechnic)

This study aims to determine the extent of the influence of e-service quality and e-WOM on customer satisfaction with the Shopee application. The population in this study is 92 students from the Business Administration Department at Bali State Polytechnic who use the Shopee application for shopping, especially for fashion items, with an unknown total number of users. The method used in this research is quantitative, with a Total Sampling technique. Hypothesis testing in this study employs statistical model analysis consisting of validity tests, reliability tests, classical assumption tests, multiple linear regression tests, T-tests, F-tests, and R² determination coefficients using SPSS22 software. The results show that there is an influence between the independent variables, namely e-service quality and e-WOM, on the dependent variable, customer satisfaction.

The Influence of Customer Participation Towards Repurchase Intention with Online Brand Community Trust and Positive eWOM as Intervening Variables in BeautyHaul

E-commerce technology is advancing rapidly in Indonesia's personal care and cosmetics industries. BeautyHaul, the first personal care e-commerce company in Indonesia, aims to strengthen relationships with customers and increase their loyalty through consumer involvement in brand communities and word-of-mouth (eWOM) marketing which have the potential to increase repurchase intention. Data were collected from a sample survey of 155 respondents using purposive sampling method, which were analysed through Partial Least Squares-Structural Equation Modelling (PLS-SEM) and PLS-Predict. The coefficient of determination R2 indicates the influence of independent variables on repurchase intention is 44%. The analysis findings show that customer participation, positive electronic word-of-mouth (eWOM) and online brand community trust has positive and significant influence on repurchase intention on direct effect. The result shows that positive electronic word-of-mouth (eWOM) and online brand community trust also have positive and significant influence as mediators between customer participation and repurchase intention. Based on the findings of this study, BeautyHaul should concentrate more on boosting customer participation in order to strengthen repurchase intention.

Deep Learning Approaches for Power Prediction in Wind–Solar Tower Systems

Wind–solar towers are a relatively new method of capturing renewable energy from solar and wind power. Solar radiation is collected and heated air is forced to move through the tower. The thermal updraft propels a wind turbine to generate electricity. Furthermore, the top of the tower’s vortex generators produces a pressure differential, which intensifies the updraft. Data were gathered from a wind–solar tower system prototype developed and established at Kyushu University in Japan. Aiming to predict the power output of the system, while knowing a set of features, the data were evaluated and utilized to build a regression model. Sensitivity analysis guided the feature selection process. Several machine learning models were utilized in this study, and the most appropriate model was chosen based on prediction quality and temporal criteria. We started with a simple linear regression model but it was inaccurate. By adding some non-linearity through using polynomial regression of the second order, the accuracy increased considerably sufficiently. Moreover, deep neural networks were trained and tested to enhance the power prediction performance. These networks performed very well, having the most powerful prediction capabilities, with a coefficient of determination R2=0.99734 after hyper-parameter tuning. A 1-D convolutional neural network achieved less accuracy with R2=0.99647, but is still considered a competitive model. A reduced model was introduced trading off some accuracy (R2=0.9916) for significantly reduced data collection requirements and effort.

Hybrid machine learning models to predict the shear strength of discontinuities with different joint wall compressive strength

ABSTRACT The shear strength of discontinuities with different joint wall compressive strength (DDJCS) is an important mechanical property. In this paper, the prediction of shear strength of DDJCS is performed using a hybrid machine learning model of artificial neural network (ANN) model with particle swarm optimisation (PSO). To develop the proposed model, a total of 168 cases are collected. Two classical models are introduced, and four performance indexes (i.e. coefficient of determination R2, mean absolute error MAE, root mean square error RMSE, and variance account for VAF) are utilised to evaluate the comprehensive predictive performances of these models. Finally, a graphical user interface (GUI) for predicting the shear strength of DDJCS is developed. The comparison results demonstrate that compared to the remaining models, this hybrid model has a better predictive performance and generalisation capability, with R2, RMSE, VAF, and MAE values of 0.999, 0.02, 99.9%, and 0.01, respectively. GUI signifies that this novel model provides an intuitive, efficient, and visual production experience for researchers. These findings can provide valuable insights for developing accurate predictive models.