Sepsis and multi-organ dysfunction syndrome (MODS) are critical clinical syndromes that need to be predicted in time to provide early clinical intervention. In this work, six machine learning models were trained to predict patient mortality: Decision Tree, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Logistic Regression, Random Forest, and XGBoost. An organized clinical dataset was pre-processed with feature selection and normalization, and then applied to train and test the models. Accuracy, Precision, Recall, and F1-score were used as evaluation metrics. Logistic Regression was the most successful overall classifier, with the best results, and KNN and XGBoost showed similar, balanced results. Conversely, Random Forest was overfitting, and Decision Tree had poor generalization. These results show that classical machine learning models, specifically Logistic Regression, KNN, and XGBoost, may offer balanced and consistent performance in sepsis mortality prediction, but the selection and tuning of models should be carefully considered for applicability in clinical contexts.

Energy management is a crucial component of smart buildings and sustainable development, enabling the efficient use of resources. Accurate forecasting of appliance-level energy demand is crucial for facilitating optimal operation and planning. This research compares the effectiveness of five ML models (LightGBM, CatBoost, XGBoost, AdaBoost, and RF) for predicting appliance energy usage. Three performance criteria are utilized to evaluate the model performance, including MAE (Mean Absolute Error), RMSE (Root Mean Square Error), and R2 (coefficient of determination). Results show that LightGBM is the most suitable algorithm, with the lowest RMSE (9.83) and the highest R2 (0.9883), followed by CatBoost and XGBoost. AdaBoost and Random Forest yielded less accurate predictions. Finally, it can be concluded that LightGBM is an effective ML model for predicting appliance-level electricity consumption, which is beneficial for energy management in smart buildings.

Cardiovascular diseases (CVD) are one of the leading causes of worldwide fatalities. A key element that contributes to these fatalities is the lag in identifying the precondition effectively, which hampers proper treatment. Early detection and accurate forecasting of CVD, combined with appropriate treatment, are crucial for safeguarding a patient’s life. In this work, performance analysis of various machine learning models (random forest (RF), adaptive boosting (ADABoost), categorical boosting (CATBoost), and extreme gradient boosting (XGBoost), is assessed on a comprehensive clinical dataset. The dataset comprises 11 attributes and 1 output attribute for 918 patients. This study adhered to a systematic process. The data underwent pre-processing first. Subsequently, performance was improved with the application of an ANOVA-based feature selection method and optimization strategy on the dataset. Subsequently, models were trained on the updated dataset, and their performance metrics, including accuracy, precision, and recall, were assessed. The findings demonstrate that CATBoost achieves an accuracy of 83.48

This paper presents a novel approach for predicting various feedstock higher heating values (HHV) using a voting ensemble machine-learning model. The proposed model, referred to as VSGB, combines Support Vector Regression (SR), Gaussian Process Regression (GR), and Boosting (BO) using a weighted sum technique. The Invasive Weed Optimization (IWO) algorithm is employed to estimate hyperparameter values of the VSGB model. Moreover, comparative performance analysis is conducted using several models, such as linear regression (LR), generalized additive model (GAM), bagging (BAG), decision tree (DT), and neural network (NN). The simulation findings demonstrate that the VSGB has a high level of accuracy in predicting the HHV derived from biomass waste. This is evidenced by the lower Root Mean Square Error (RMSE) and Average Absolute Relative Difference (AARD%) values (0.813 and 2.827%, respectively) compared to other Machine Learning (ML) predictive models. Additionally, the present study establishes an empirical correlation between the higher heating value (HHV) and the input characteristics carbon (C), hydrogen (H), oxygen (O), nitrogen (N), and sulphur (S) through the utilization of the IWO algorithm.

The HVAC system can achieve low energy consumption as ML-based models accurately estimate the building’s energy use and load demands. Therefore, in this work, an extreme gradient boosting (XGBoost) ensemble model is proposed for predicting energy usage based on heating and cooling Loads (HL and CL). Furthermore, RF, LR, KNN, and SVR are also designed for comparison analysis. The results show that XGBoost outperforms all the applied algorithms, achieving the lowest values of RMSE (0.407 and 0.858) and MSE (0.166 and 0.737) in both cases. Furthermore, its performance is also compared with the models presented in the literature. Finally, it can be concluded that the proposed XGBoost is superior, robust, and efficient for predicting HL and CL, respectively.

The HVAC unit helps reduce overall energy consumption. ML models can enhance HVAC performance by accurately predicting a building's energy consumption and load utilization. Therefore, this study presents a stacked ensemble model that incorporates extreme gradient boosting (XGB), decision tree (DT), and Random Forest (RF) algorithms to predict the energy consumption of heating and cooling loads (HL and CL) in buildings. The performance of the proposed stacked ensemble is compared to other machine learning predictive models such as Ridge, Lasso, K Nearest Neighbor (KNN), Support Vector Regression (SVR), and Artificial Neural Network (ANN). Bayesian optimization is used to determine the hyperparameter values of the ML algorithms. The results show that the proposed predictive model has the lowest root mean square value (RMSE) of 0.484 and 0.948 for HL and CL, respectively, compared to other machine learning models. Additionally, the efficacy of the stack model is evaluated using a time series dataset about HVAC energy consumption in residential buildings. The simulation results indicate that the stack model outperformed the other prediction models, achieving a root mean square error (RMSE) of 0.1810. In conclusion, the proposed predictive model is more efficient than traditional models in forecasting energy consumption by HL, CL, and HVAC systems.

In this article, a new control algorithm is proposed for the frequency control of a fuel cell-powered nano grid. The proposed control algorithm amalgamates the 2 degrees of freedom PD (2 PD) and proportional-integral-derivative (PID) control schemes leading to 2 PD-PID. The optimal values of design parameters of 2 PD-PID controllers are estimated using different optimization techniques. The convergence performance of particle swarm optimization (PSO) is found to be better compared to Gorilla Troop Optimizer (GTO), African Vultures Optimization Algorithms (AVOA), and Geometric Mean Optimizer (GMO). The PI, PID, 2 degrees of freedom PI (2PI), and 2 degrees of freedom PID (2 PID) control schemes are also designed for comparative analysis. Results show that the proposed control scheme attains a lower settling time value (1.2374e-06 s) and peak time (1.2262e-06 s) than PI, PID, 2 PI, and 2 PID. Furthermore, 2 PD-PID control schemes improve the IAE and ISE at 86%-98% and 83%-97% with respect to other control schemes for continuous change in the load conditions. Finally, it can be interpreted that the 2 PD-PID control scheme is robust, efficient, and effective compared to other designed control schemes.

This work incorporates an adaptive learning-based boosting (ADboost) ML classifier to classify four types of fuel: agricultural residue, coals, wood, and produced biomass. Further, the ADboost’s hyperparameters, such as learning rate, maximum number of splits, and minimum leaf size are adjusted using teaching learning-based optimization (TLBO), resulting in TADboost. The performance of TADboost is compared against various popular ML (NN, BAG, NB, and SVM) models. Simulated result reveals that the suggested classifiers outperform other compared ML classifiers for fuel classification with the classification accuracy, precision, recall, F1-score, and kappa as 0.9659, 0.9671, 0.9449, 0.9558, and 0.9482, respectively.

The gas turbine in a combined cyclic power plant (CCPP) produces harmful gases like carbon monoxide (CO) and nitrogen oxide (NOx) into the atmosphere. It is evident to monitor the rate at which these gases are produced during power generation to comply with the industrial standard for emission. Therefore, a system is required to continuously monitor the emission from the CCPP gas turbine. Hence, this work aims to design a stacked ensemble machine learning (SEM) based predictive model for CO and NOx emission from a CCPP gas turbine. The neural network for regression (NNR), a generalized additive model (GAM), and the bagging of regression trees (BT) act as the base learners. A generalized regression neural network (GRNN) is used as a meta-learner for SEM. The hyperparameters of SEM are optimized using a Bayesian optimization algorithm for CO and NOX prediction. In addition to this, the performance of SEM is compared with support vector regression (SVR), decision tree (DRT), and linear regression (LIR). Simulation results demonstrate that SEM can reduce the RMSE 5.7–93.8% for NOx and 1%-41.5% for CO compared to other ML techniques. Finally, comparing the results with ML techniques existing in the literature shows the higher predictive accuracy of the proposed SEM.

Differential evolution (DE) is a practical evolutionary algorithm (EA) widely employed for addressing continuous optimization problems. Opposition-based learning (OBL) emerges as a potent method among the techniques enhancing EA performance. The BetaCOBL variant represents a pinnacle in this domain. However, BetaCOBL’s utilization of the promising regions of the search space remains partial, owing to its dependence on a non-adaptive framework. Consequently, its efficacy might dwindle as optimization progresses. We aimed to introduce an enhanced version of BetaCOBL, termed adaptive BetaCOBL (ABetaCOBL). ABetaCOBL commences by adapting the search space based on population distribution and subsequently identifying opposite solutions. We evaluated the efficacy of embedding ABetaCOBL into DE algorithms through experiments. Our experimental results substantiate that ABetaCOBL outperforms its precursor and resilient OBL variants (e.g., ABetaCOBL outperforms iBetaCOBL-eig in 19 out of 58 problems with NL-SHADE-LBC and in 22 out of 58 problems with NL-SHADE-RSP).

This work proposes a new blended stacked ensemble machine-learning model (BEM) to predict biomass's higher heating value (HHV) from the ultimate analysis. Gorilla troop optimization (GTO) is utilized to estimate the hyperparameter values of BEM, leading to GBEM. In GBEM, support vector regression (SUVR), Gaussian process regression (GAPR), and Decision Tree (DETR) are used as the base learner, whereas adaptive linear neural network (ADALINE) is used as a meta-learner, respectively. Furthermore, Linear Regression (LIR), generalized additive model (GEAM), and bagging of regression trees (BAGG) are also designed for comparison purposes. Results reveal that GBEM predicts the HHV with a lower AARD% (2.959%) value than other designed ML predictive models. In addition to this, a predictive equation that gives the relationship between HHV and the ultimate analysis parameters C, H, O, N, and S is also derived using GTO.

The construction industry consumes 35% of all global energy. Building energy conservation is critical for lowering emissions and consumption. Properly functioning the building's heating, ventilation, and air conditioning (HVAC) unit helps to reduce energy consumption. Predicting building energy consumption with machine learning (ML) models can help to improve HVAC functionality. As a result, the performance of various ML predictive models based on k-nearest neighbor (KNN), artificial neural network (ANN), support vector regression (SVR), and Ridge and Lasso regression models is investigated in this work for the prediction of energy usage. Furthermore, Bayesian optimization for different random states (RS) is used to estimate the hyperparameters of the ML models that have been implemented. The results show that ANN performs best for RS values between 0 and 75. However, SVR achieves the lowest RMSE for RS, equal to 25, 50, 100, 150, and 200, compared to ANN, KNN, Ridge, and Lasso (RMSE=2.910), respectively. Finally, SVR predicts energy consumption more accurately than other designed models in most cases.

In this work, a higher-order Proton Exchange Membrane (PEM) Fuel Cell system is controlled using a cascaded control approach. The primary and secondary controllers, NPID/PI, are non-linear proportional-integral-derivative and proportional-integral, respectively. The suggested cascade approach controls the stack voltage by adjusting the air compressor voltage to maintain the oxygen excess ratio value within limits. Two additional cascade control structures, PID/PI and FOPID/FOPI (fractional order PID and fractional order PI), are also created for a fair comparison. A genetic algorithm is used to determine the controller's optimal parameters by minimising the time integral absolute error of the primary controller. Results show that NPID/PI control structure achieves the minimum value of settling time and overshot (0.9514 s, 0.021%) compared to FOPID/PI (1.8980 s, 0.036%) and PID/PI (3.2308 s, 0.092%), respectively. Finally, it can be concluded from the result of setpoint tracking, disturbance rejection, and noise suppression that the proposed controller is efficient and robust compared to other designed controllers. (c) 2023 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

This article presents a new cascaded control strategy to control the power flow in a renewable-energy-based microgrid operating in grid-connected mode. The microgrid model is composed of an AC utility grid interfaced with a multi-functional grid interactive converter (MF-GIC) acting as a grid-forming converter, a photovoltaic (PV) power-generation system acting as grid-feeding distributed generation unit, and various sensitive/non-sensitive customer loads. The proposed control strategy consists of a fractional order PI (FO-PI) controller to smoothly regulate the power flow between the utility grid, distributed generation unit, and the customers. The proposed controller exploits the advantages of FO (Fractional Order) calculus in improving the steady-state and dynamic performance of the renewable-energy-based microgrid under various operating conditions and during system uncertainties. To tune the control parameters of the proposed controller, a recently developed evaporation-rate-based water-cycle algorithm (ERWCA) is utilized. The performance of the proposed control strategy is tested under various operating conditions to show its efficacy over the conventional controller. The result shows that the proposed controller is effective and robust in maintaining all the system parameters within limits under all operating conditions, including system uncertainties.

A figurative language expression known as sarcasm implies the complete contrast of what is being stated with what is meant, with the latter usually being rather or extremely offensive, meant to offend or humiliate someone. In routine conversations on social media websites, sarcasm is frequently utilized. Sentiment analysis procedures are prone to errors because sarcasm can change a statement’s meaning. Analytic accuracy apprehension has increased as automatic social networking analysis tools have grown. According to preliminary studies, the accuracy of computerized sentiment analysis has been dramatically decreased by sarcastic remarks alone. Sarcastic expressions also affect automatic false news identification and cause false positives. Because sarcastic comments are inherently ambiguous, identifying sarcasm may be difficult. Different individual NLP strategies have been proposed in the past. However, each methodology has text contexts and vicinity restrictions. The methods are unable to manage various kinds of content. This study suggests a unique ensemble approach based on text embedding that includes fuzzy evolutionary logic at the top layer. This approach involves applying fuzzy logic to ensemble embeddings from the Word2Vec, GloVe, and BERT models before making the final classification. The three models’ weights assigned to the probability are used to categorize objects using the fuzzy layer. The suggested model was validated on the following social media datasets: the Headlines dataset, the “Self-Annotated Reddit Corpus” (SARC), and the Twitter app dataset. Accuracies of 90.81%, 85.38%, and 86.80%, respectively, were achieved. The accuracy metrics were more accurate than those of earlier state-of-the-art models.

Chemotherapy is a widely used cancer treatment method globally. However, cancer cells can develop resistance towards single-drug-based chemotherapy if it is infused for extended periods, resulting in treatment failure in many cases. To address this issue, oncologists have progressed towards using multi-drug chemotherapy (MDC). This method considers different drug concentrations for cancer treatment, but choosing incorrect drug concentrations can adversely affect the patient's body. Therefore, it is crucial to recognize the trade-off between drug concentrations and their adverse effects. To address this issue, a closed-loop multi-drug scheduling based on Fractional Order Internal-Model-Control Proportional Integral (IMC-FOPI) Control is proposed. The proposed scheme combines the benefits of fractional PI and internal model controllers. Additionally, the parameters of IMC-FOPI are optimally tuned using a random walk-based Moth-flame optimization. The performance of the proposed controller is compared with PI and Two degrees of freedom PI (2PI) controllers for drug concentration control at the tumor site. The results reveal that the proposed control scheme improves the settling time by 43% and 21% for V-X, 54% and 48 % for V-Y, and 48% and 40% for V-Z, respectively, compared to PI and 2PI. Therefore, it can be concluded that the proposed control scheme is more efficient in scheduling multi-drug than conventional controllers.

The Kraft recovery process in the paper mills is highly complex and nonlinear. The evaporation of black liquor using Multiple Effect Evaporators (MEE) in the recovery unit of the Kraft process is an energy-consuming procedure. The challenge lies in designing a proficient control strategy to conserve energy and guarantee good product quality. In this article, the 2-DOF-PI controller is designed to enable efficient control of MEE and eliminate process uncertainties. The Heptads’ Effect Falling Film Evaporator in the backward feed flow configuration is used as a working platform. The steady-state and transient behavior of MEE is modeled and analyzed to assist in controller design. The steady-state process parameters, that guarantee optimum energy efficiency, are estimated using Moth–Flame Optimization (MFO). MFO is also employed to estimate the optimal 2-DOF-PI controller parameters to attain improved product quality and energy-efficient performance of MEE. The competence of MFO towards controller tuning is shown by a fair comparison with some well-known optimization techniques. The quantitative investigation of the statistical results demonstrates that MFO outperforms the other algorithms. To validate the proficiency of the designed control scheme, its performance is compared with basic PI/PID control strategy for set-point tracking, noise suppression, and process uncertainties. The simulation results indicate that the designed MFO-tuned 2-DOF-PI controller offers efficient control action and, therefore, proves to be an appropriate control algorithm to ensure sustainable production.

A combined cycle power plant (CCPP) employs gas and steam turbines to generate 50% more power while utilizing the same fuel as a normal single cycle plant. The performance of a CCPP under full load is affected by a variety of factors such as weather, process interactions, and coupling, which makes it challenging to operate. Therefore, a reliable assessment of the maximum output power of a CCPP is required to improve plant reliability and monetary performance. In this paper, a predictive model based on a generalized additive model (GAM) is proposed for the electrical power prediction of a CCPP at full load. In GAM, a boosted tree and gradient boosting algorithm are considered as shape function and learning technique for modeling a non-linear relationship between input and output attributes. Furthermore, predictive models based on linear regression (LR), Gaussian process regression (GPR), multilayer perceptron neural network (MLP), support vector regression (SVR), decision tree (DT), and bootstrap-aggregated tree (BBT) are also designed for comparison purposes. Results reveal that GAM improves the RMSE by 74%, 68.8%, 70.3%, 54.8%, 21.2%, and 17.3% compared to LR, GPR, MLP, SVR, DT, and BBT, respectively. Furthermore, the results of the Man-Whitney U test and rank analysis also confirm the effectiveness of GAM for energy prediction of CCPP. Finally, it can be concluded that the proposed method is effective, robust, and accurate for the assessment of the maximum output power of a CCPP to improve plant consistency and financial performance.

The concept of distributed generators (DG) and their control has been evolved as a key area of research, to ensure the sustainability of a microgrid (MG). Design and implementation of a proactive cascaded control strategy is an effective method to make the MG more sustainable and resilient towards uncertainties. In this paper, a novel cascaded control strategy consisting of two degree of freedom (2DOF) PI and an internal model controller (IMC) is proposed to effectively control the grid interactive converter (GIC). The proposed control strategy exploits the combined benefits of 2DOF-PI and IMC in improving the steady-state and dynamic performance of the GIC. The primary function of the proposed control strategy is to effectively control the GIC to facilitate smooth power flow between MG and utility. In addition to this, the proposed control strategy enables the GIC to offer different ancillary services that include unbalanced load current compensation, reactive power compensation, and harmonic current reduction. To achieve this multi-functional feature, appropriate reference currents are extracted, and a control strategy is implemented in a synchronous reference (dq0) frame. An evaporation rate-based water cycle algorithm (ERWCA) optimization technique is employed to estimate the optimal design parameters of the cascaded controller. The closed-loop stability of the system is verified using frequency and time domain analyses for the estimated optimal design parameters. To show the effectiveness of the proposed control strategy, various case studies are considered, and results are compared with existing methods.

Accurate building energy consumption prediction is essential for achieving energy savings and boosting the HVAC system's efficiency of operations. Therefore, in this work, a novel ensemble predictive model, which combines the weighted linear aggregation of Gaussian process regression (GPR) and least squared boosted regression trees (LSB), leading to WGPRLSB, is proposed for the accurate estimation of energy usage in the cases of Heating Load (HL) and Cooling Load (CL). Marine predator optimization (MPO) is used to evaluate the optimal values of the design parameters of the proposed methodology. Further, predictive models based on linear regression (LR), support vector regression (SVR), multilayer perceptron neural network (MLPNN), decision tree (DT), and generalized additive model (GAM) are also designed for comparison purposes. The results reveal that the value of RMSE is reduced by 12.4%–70.7% (HL) and 39.7%–64.9% (CL) for WGPRLSB in comparison to the other predictive models. The results of the performance index (PI) also confirm the effectiveness of the proposed model energy consumption prediction for HL and CL. Furthermore, the performance investigation on the second dataset reveals that WGPRLSB achieves the highest value of VAF (97.20%) compared to other designed models. It may be concluded that the proposed WGPRLSB accurately forecasts building energy demands.