Average Absolute Percentage Error Research Articles

State of health estimation for lithium-ion battery based on particle swarm optimization algorithm and extreme learning machine

Lithium-ion battery State of Health (SOH) estimation is an essential issue in battery management systems. In order to better estimate battery SOH, Extreme Learning Machine (ELM) is used to establish a model to estimate lithium-ion battery SOH. The Swarm Optimization algorithm (PSO) is used to automatically adjust and optimize the parameters of ELM to improve estimation accuracy. Firstly, collect cyclic aging data of the battery and extract five characteristic quantities related to battery capacity from the battery charging curve and increment capacity curve. Use Grey Relation Analysis (GRA) method to analyze the correlation between battery capacity and five characteristic quantities. Then, an ELM is used to build the capacity estimation model of the lithium-ion battery based on five characteristics, and a PSO is introduced to optimize the parameters of the capacity estimation model. The proposed method is validated by the degradation experiment of the lithium-ion battery under different conditions. The results show that the battery capacity estimation model based on ELM and PSO has better accuracy and stability in capacity estimation, and the average absolute percentage error is less than 1%.

Mechanical safety prediction of a battery-pack system under low speed frontal impact via machine learning

In response to nonrenewable energy consumption and environmental pollution, the development of electric vehicles has accelerated. The safety of electric vehicles has garnered considerable attention. There are numerous incalculable foreign objects on the road that may collide with or scratch the battery-pack’s frontal, resulting in battery-pack system damage or even explosion. This poses a significant risk to the safety of passengers and drivers. In this paper, a diversity of mechanical safety prediction models for battery-pack systems are proposed. These models support data-driven structural optimization of the battery-pack system by utilizing the numerical results of the bottom shell deformation. These simulation-based prediction models combine response surface method and machine learning algorithms. First, a nonlinear finite element model of the battery-pack system is established. The efficacy of the model is verified using constrained modal analysis in a variety of commercial software packages. Second, the collision simulations are executed and the data in various collision conditions are collected. Different sample sizes are used to develop various response surface models and machine learning models. The machine learning algorithms adopt support vector machine, Gaussian process regression, and neural network models. Third, the prediction accuracy of multiple prediction models is investigated according to error functions. The results show that the neural network model can predict the most accurate deformation under the condition of low speed frontal impact. The prediction average absolute percentage error of the neural network model is only 0.34% within the design domain, and only 2.54% outside the design domain. The proposed prediction model can be used for reliable design of the battery-pack system in electric vehicles. It also can be employed to design the early warning system of the battery-packs.

Passenger Flow Prediction Method based on Hybrid Algorithm: Intelligent Transportation System

Forecasting passenger flow at metro transit stations is a useful method for optimizing the organization of passenger transportation and enhancing operational safety and transportation efficiency. Aiming at the problem that the traditional ARIMA model has poor performance in predicting passenger flow, a hybrid prediction method based on ARIMA-Kalman filtering is proposed. In this regard, ARIMA model training experimental samples are integrated with Kalman filter to create a prediction recursion equation, which is then utilized to estimate passenger flow. The simulation experiment results based on the inbound passenger flow data of Nanjing metro station show that compared with the single ARIMA model, the root mean square error of the prediction results of the proposed ARIMA-Kalman filter hybrid algorithm is reduced by 257.106, and the mean absolute error decreased by 145. 675, the mean absolute percentage error dropped by 5. 655%, proving that the proposed hybrid algorithm has higher prediction accuracy. The experiment results based on the passenger flow data of Nanjing metro station show that compared to a single ARIMA model, the proposed ARIMA Kalman filtering hybrid algorithm reduces the root mean square error of the prediction results by 257.106, the average absolute error by 145.675, and the average absolute percentage error by 5.655%. It has been proven that the proposed hybrid algorithm has higher prediction accuracy.

PREDICTION OF PARAMETERS OF BOILER SUPERHEATER BASED ON TRANSFER LEARNING METHOD

The superheater in the boiler is the key of equipment connecting high-temperature steam to the turbine for power generation. At present, the problems of large variable fluctuations, strong timing coupling, and multi-power plant data utilization prevent the temperature, flow, and pressure prediction of the boiler superheater. In this paper, a method for predicting the parameters of boiler superheater based on a transfer learning model is proposed, which realizes the joint utilization of data from multiple power plants. The method first collects data from a waste incineration boiler power plant for pre-training the long short-term memory (LSTM)-transformer model, and then completes the transfer learning training on the new power plant. The proposed method has the advantages of high prediction accuracy, good robustness, and more reliable location prediction with drastic changes. The predictions on the test set are within ± 5% of the experimental value. Compared with the model not trained by the transfer learning, the proposed method achieves the lowest relative errors for all prediction intervals in the 3-15 min range. Compared to the linear regression (LR), support vector regression (SVR), and random forest (RF), the proposed method improves the average absolute percentage error (MAPE) by 30%, 13%, and 20%, respectively. Flatter loss sharpness value and better robust performance obtained from the transfer learning method is verified by an experimental verification. Finally, a digital system design for power plants with real-time data visualization monitoring, parameter prediction, and fault warning functions are implemented.

WSN-IoT Forecast: Wireless Sensor Network Throughput Prediction Framework in Multimedia Internet of Things

Accurate throughput predictions can significantly improve the quality of experience (QoE), where QoE denotes a network’s capacity to provide satisfactory service. By increasing the results of good throughput predictions, the best strategy can be planned for managing data transmission networks with the aim of better and faster data transmission, thereby increasing QoE. Consequently, this paper investigates how to predict the throughput of wireless sensor networks utilizing multimedia data. First, we conducted a comparative analysis of relevant prior research on the topic of throughput prediction in Multimedia Internet of Things (Multimedia IoT). We developed a throughput prediction framework for wireless sensor networks based on what we learned from these studies using machine learning. The Throughput Prediction Framework identifies historical throughput data and employs these traits to predict throughput. In the final phase, multiple camera nodes and local servers are utilized to test a framework for throughput prediction. Our analysis demonstrates that WSN-IoT predictions are quite precise. For a 1-second time breakdown, the average absolute percentage error for all investigated scenarios ranges from 1 to 8 percent.

การพยากรณ์ยอดขายและความต้องการสินค้าในอุตสาหกรรมผลิตอุปกรณ์ไฟฟ้า กรณีศึกษา บริษัทผลิตอุปกรณ์ไฟฟ้าแรงสูง

The purpose of this study is to find the product demand forecasting model from the artificial neural network method in a case study of a high-voltage equipment manufacturing company. And reduce fines for late delivery of goods by using the Matlab program. The forecasting model is established by using time series neural network method, and the nonlinear autoregressive neural network (NARX) function model is established by combining with external input data. Then the forecast deviation is the average deviation percentage error (MAPE) uses the least dynamics model. By using data that collects product sales from 2019 – 2022 total 48 set of data are divided into 2 data sets used in the study. by the first dataset is learning 36 datasets and the second dataset is learning 12 datasets. The model of this study was defined with the number of hidden layer nodes as 2, 6 and 10. The time delay was 2 and 3. Forecasting models are combined into 6 models. The best model result of each item from NARX forecast through the learning process Levenberg-Marquardt (trainlm), including models 1, 4, 5, 4, 3, 3, and 2 sequential The results show the least average absolute percentage error at 8.32, 5.85, 2.61, 3.39, 2.91, 2.19 and 7.76 and the results of the forecasting model were analyzed to reduce fines from selling all 7 items Able to reduce penalties for late delivery of 4 items showing the percentage of difference in cost reduction equal to 0.28% 0.24% 0.32% and 0.22%. It can be concluded that the prediction of the neural network inverted neural network model It can be used to forecast product sales and product demand as well. And reduce the cost of late delivery fines for better results.

A voltage control method for distribution networks based on TCN and MPGA under cloud edge collaborative architecture

A distribution network voltage control method based on TCN and MPGA under cloud edge collaborative architecture is proposed to address the issues of heavy computational burden and inability to balance economic efficiency in traditional decentralized and centralized voltage control methods. Firstly, under the cloud edge collaboration architecture, the power prediction model based on time convolutional networks is trained in the cloud, and the edge predicts the load demand and new energy active output in each region based on the TCN Attention power prediction model issued by the cloud. Then, considering safety and economy comprehensively, with the goal of minimizing the daily network loss and voltage regulation cost in the distribution network, the Pareto solution set for the optimal operation point of each voltage regulation equipment in the entire distribution network was solved based on an improved multi population genetic algorithm at the distribution cloud main station. Finally, the optimal distribution network voltage control scheme is obtained based on the Pareto solution set. Based on the simulation system, experimental testing was conducted on the proposed method, and the results showed that its daily voltage regulation cost and network loss were 863.4 yuan and 43.73 MW h, respectively. The average absolute percentage error of the prediction model was 4.51 %, which was superior to other comparative methods.

Development and comparison of different artificial intelligence-based models for viscosity of in-situ cross-linked acids

The use of in-situ cross-linked acids (ICAs) is common in stimulating oil and gas reservoirs. However, their rheological models have not been widely incorporated into simulation software due to their complex behavior and lack of valid data. Accordingly, in this study, development of new models for estimation of viscosity of ICAs as a function of pH and shear rate was intended. For this purpose, 33 experimental data points were collected from literature and methods of genetic programming, neural network (NN), and fuzzy logic (FL) were used to develop models. In summary, all three models performed equally well and better than a previously published model, with an R 2 = 0.99 and an average absolute percent error of 7%. In terms of computational costs, genetic programming correlation was found to be 76 and 786% faster than NN and FL, respectively. Therefore, the model developed by genetic programming was suggested to be used in numerical solvers for estimation the viscosity of ICAs as a function of pH and shear. Sensitivity analysis on temperature, pH, and shear rate showed that pH would have the highest impact on the apparent viscosity of the considered ICAs.

A novel interval-valued carbon price analysis and forecasting system based on multi-objective ensemble strategy for carbon trading market

Accurate forecasting of carbon price is crucial for the efficient management and stable operation of carbon markets. Earlier studies are limited to point and interval forecasts based on single-valued carbon price and lack analysis and forecasting based on interval-valued carbon price. Therefore, this study proposes a novel analysis and forecasting system from a new perspective of interval-valued carbon price. Specifically, a carbon price analysis sub-system is developed to investigate the directional causal relationship between the upper and lower bounds of the interval-valued carbon price series. The carbon price forecasting sub-system is developed by designing a data preprocessing module, sub-model forecasting module, and multi-objective ensemble module. The data preprocessing module adopts the decomposition algorithm to preprocess the interval-valued carbon price. Then the sub-model forecasting module utilizes multiple neural network models to predict the highest and lowest prices. Finally, the multi-objective ensemble module adopts a non-linear and multi-objective ensemble strategy to ensemble the forecasting results of the sub-models. It can be found that the consideration of both upper and lower bounds of interval-valued carbon price within the range leads to higher prediction accuracy for the highest or lowest price predictions. Additionally, the ensemble model can effectively leverage the strengths of individual sub-models, resulting in more precise and stable predictions. The average absolute percentage errors for the highest and lowest price predictions in the Hubei and Guangzhou carbon trading markets are 0.8574%, 1.2738%, 0.9774%, and 1.8217% respectively, vividly demonstrating the effectiveness of the proposed system in carbon price prediction.

Reliability and Validity of the Repetition-Counting Feature of the Push Band 2.0 at Different Repetition Speeds

Purpose: To assess the interdevice reliability and validity of the repetition-counting feature of the Push Band 2.0. Methods: Thirty college-aged participants (aged 18–24 years) simultaneously wore two Push Band 2.0 devices and performed 10 common resistance training exercises at four different tempos over the course of two testing sessions. Twelve repetitions were completed with visual confirmation for each set of exercises and compared with repetition estimates from the Push Band 2.0. Interdevice reliability was quantified using single measures intraclass correlation coefficients with 95% confidence intervals while validity was assessed via mean absolute percent error and mean percent error. Results: Interdevice reliability was found to be good to very good regardless of exercise type or tempo, as all intraclass correlation coefficients were >.770. Validity of the repetition-counting feature of the device was dependent on both exercise type and tempo, as exercises that did not involve rotation of the device throughout the movement demonstrated greater mean absolute percent error (31.0% average of all four tempos) and mean percent error (−29.9% average of all four tempos) than those that required such rotation (average mean absolute percent error of 3.5% and mean percent error of −1.6% across all four tempos). Conclusions: This study supports the reliability of the repetition-counting feature of the Push Band 2.0. However, device accuracy appears to be dependent on the type of movement and the speed at which the movement is performed, with greater accuracy observed during faster exercise tempos and exercises involving rotation of the device during movement execution.

Bayesian inference for parameter identification in mechanistic models, exemplified using a cow lifetime performance model

Mechanistic models are valuable tools for studying the underlying mechanisms of complex biological phenomena. For example, cow lifespan models can be used to identify differences in resource acquisition and allocation strategies between individuals, which is relevant for decision-making in breeding programs. In such models, differences in simulated traits between individuals are consequences of the parameter set that represents the genetic potential of each animal and its interaction with the environment. This indicates that the identification of these differences is essentially a search for individual parameters. In mechanistic models, this search is generally a non-convex problem that has different local minima because the parameters interact within these models. Due to this and to the simulation time length (e.g. years), there is uncertainty associated with the inference of the parameter values for each individual. This uncertainty can be quantified using Bayesian inference since this approach treats the model parameters as random variables with an underlying probability distribution that describes them. The objective of this work was to employ the Delayed Rejection Adaptive Metropolis (DRAM) algorithm to identify the parameters of a cows’ lifespan model using two datasets of Holstein cows. The datasets contain periodic measurements of Milk Yield (MY), BW, and Body Condition Score (BCS). Additionally, one of the two datasets has information of BW from birth to first calving. The average Mean Absolute Percentage Error (MAPE) minimisation between the simulated and experimental data (MY, BW and BCS) was used as the objective function for parameter search. The Bayesian inference performance was compared with four optimisation metaheuristic approaches: Differential Evolution, Genetic Algorithm, Particle Swarm Optimisation, and Simulated Annealing. Although the results show that all methods are efficient in finding parameter values that reduce the distance between the simulated and experimental data (MAPE < 10%), the DRAM method is more efficient in terms of computational cost, and the parameter distributions obtained with this method offer more information about the statistical properties of each parameter (e.g. median).

Research on the Implementation Path of Tourism Contributing to the Development of Rural Revitalization Combined with Elastic Regression Network Algorithm

Abstract This paper analyzes the power mechanism of tourism to help rural revitalization by elaborating the core content of rural tourism and rural revitalization. It also combines the advantages of the Lasso regression method and ridge regression method to form a flexible regression network model, which is used to predict the amount of tourists and tourists’ demand for rural tourism so as to put forward the implementation path of the tourism industry to help rural revitalization development. To verify the effectiveness of the proposed algorithm, data analysis is carried out in rural areas of L city as an example. The results show that compared with the Lasso regression model, the average absolute percentage error decreased by 2.41%, the average absolute error decreased by 2,233, and the root-mean-square error decreased by 2,967, which indicates that the elasticity regression network prediction model has a stronger generalization ability and a better prediction ability, and it can predict the number of tourists more accurately, and it can provide data for the tourism industry to contribute to rural revitalization and development. Reference.

ANN based-model for estimating the boron permeability coefficient as boric acid in SWRO desalination plants using ensemble-based machine learning

This study examines the relationship between input conditions and the prediction of boron rejection in full-scale seawater reverse osmosis (SWRO) desalination plants using ensemble-based machine learning. While reverse osmosis is the dominant desalination technology, limited research has focused on analyzing plant performance under actual operating conditions. To address this gap, we developed and implemented machine learning algorithms to forecast boron permeability coefficient values, which are indicative of boron rejection concentrations in the permeate. Our analysis utilizes data from a SWRO desalination plant in southeast Spain, examining various input variables and their influence on the prediction of these parameters. The results demonstrate that our ensemble-based machine learning approach can predict boron permeability coefficient values with a reasonable margin of error of 1 mgL−1, as evidenced by mean average error (MAE) and mean absolute percentage error (MAPE) values of 7.93·10–8 and 11.8 %, respectively. In conclusion, an innovative application of artificial intelligence algorithms in the field of water purification under real operational conditions has been introduced, thus introducing valuable insights into the use of machine learning algorithms for forecasting boron rejection concentrations in full-scale SWRO desalination plants. The findings lay the foundation for future researches exploring automated and deep-learning methods in water purification.

Development of a dynamic demand response quantification and control framework for fan-coil air-conditioning systems based on prediction models

The air-conditioning demand response has gradually become an effective approach to reduce peak load of the national grid. However, with the increasing penetration of renewable energy generation sources, static performance evaluation of demand response that only fits one certain scenario cannot meet the demandsofthe times. To address this issue, a novel dynamic demand response quantification and control framework was built based on the secondary development of building energy simulation software and multilayer perceptron algorithm. In this framework, the demand response evaluation method can be used for different global temperature adjustment strategies under different external disturbances and internal loads, and the dynamic demand response control method could be further adopted for load reduction prediction and setpoint adjustment recommendation. The performance prediction model based on artificial neural network can predict the maximum load shedding during demand response period with an average absolute percentage error of 15 %, and the prediction error of load shedding intensity is ± 1 W/m2 for over 80 % of samples. Besides, the setpoint adjustment prediction model achieved an average absolute percentage error of 3 % while its prediction error is ± 0.25 °C for over 99 % of samples. The results prove that this research can provide reasonable suggestions for customers and load aggregators to achieve accurate control of setpoint adjustment in dynamic demand response conditions.

Impact of information on the number of traffic accidents on the outcome of the forecast

Every year, more and more vehicles appear on the world's roads. This leads to increased traffic on the roads. Road accidents have become a rapidly growing threat. They cause loss of human life and economic assets. This is due to the rapid growth of the world's human population and the very rapid development of motorization. The main problem in forecasting and analyzing data on the number of traffic accidents is the small size of the dataset that can be used for analysis in this regard. And on the other hand, road accidents cause, globally, millions of deaths and injuries annually is their density in time and space. It is worth noting that the pandemic has reduced the number of traffic accidents. However, the value is still very high.&#x0D; The purpose of the article is to assess the impact of information on the number of traffic accidents on the outcome of the forecast. To this end, using historical statistical data, the forecast of the number of traffic accidents for the following years was determined, and how this variability of the input data affects the value of the average percentage error of the forecast was determined. Based on the study, it can be concluded that a smaller number of input data, historical data on the number of accidents, instead of 32 years, 7 years, makes the determination of the forecast of the number of accidents for subsequent years, is at a satisfactory level, the average absolute percentage error of MAPE less than 7%. The article concludes with the determination of the forecast for future years. It is worth noting that the prevailing pandemic distorts the results obtained.

Measuring thermal comfort using wearable technology in transient conditions during office activities

This paper presents an experiment for assessing thermal comfort of occupants in the built environment, from a subjective perspective, focusing on office environment; a dedicated measurement campaign using sensors for acquisition of physiological and environmental parameters was conducted. Skin temperature was measured with two sensors: a minimally invasive sensor for measuring wrist temperature, and a thermal camera to retrieve forehead temperature; simultaneously, heart rate variability was measured using a wearable device. 15 participants were exposed to dynamic changes of air temperature. Data was collected to measure the participants’ thermal sensation vote, with machine learning algorithms. Decision Tree provided higher performances, using a dataset made of wrist temperature, heart rate variability features and air temperature, with mean average error and mean absolute percentage error of 0.86 and 20.9%. The research contributes to thermal comfort personalization in the built environment, to improve well-being and productivity of occupants using minimally invasive sensor network.

A feature fusion-based convolutional neural network for battery state-of-health estimation with mining of partial voltage curve

Accurately estimating the state-of-health of batteries is critical for effective battery monitoring and management. However, the estimation remains challenging due to dynamic operation environments and complex battery degradation patterns. In this study, a feature fusion-based convolutional neural network is proposed for battery state-of-health estimation based on voltage measurements obtained during a partial cycle. Instead of directly feeding the voltage data as input, three feature sequences are first extracted, including the capacity versus voltage curve and its differentiation with respect to voltage and life cycle. The objective is to exploit more effective information from intra-cycle and inter-cycle perspectives. Then, an element-wise addition is embedded as a feature fusion operation in the proposed convolutional neural network to generate more efficient feature maps when dealing with multiple model inputs. To validate the performance of the proposed methodology, eighteen batteries from three battery datasets are utilized for comparative studies. Experimental results demonstrate that the data preprocessing from both intra-cycle and inter-cycle perspectives, along with the adoption of the feature fusion operation, significantly improve the accuracy of battery state-of-health estimation, with an average mean absolute error and mean absolute percentage error being no more than 0.0028 and 0.32 %, respectively.

Sensorized T-Shirt with Intarsia-Knitted Conductive Textile Integrated Interconnections: Performance Assessment of Cardiac Measurements during Daily Living Activities.

The development of smart wearable solutions for monitoring daily life health status is increasingly popular, with chest straps and wristbands being predominant. This study introduces a novel sensorized T-shirt design with textile electrodes connected via a knitting technique to a Movesense device. We aimed to investigate the impact of stationary and movement actions on electrocardiography (ECG) and heart rate (HR) measurements using our sensorized T-shirt. Various activities of daily living (ADLs), including sitting, standing, walking, and mopping, were evaluated by comparing our T-shirt with a commercial chest strap. Our findings demonstrate measurement equivalence across ADLs, regardless of the sensing approach. By comparing ECG and HR measurements, we gained valuable insights into the influence of physical activity on sensorized T-shirt development for monitoring. Notably, the ECG signals exhibited remarkable similarity between our sensorized T-shirt and the chest strap, with closely aligned HR distributions during both stationary and movement actions. The average mean absolute percentage error was below 3%, affirming the agreement between the two solutions. These findings underscore the robustness and accuracy of our sensorized T-shirt in monitoring ECG and HR during diverse ADLs, emphasizing the significance of considering physical activity in cardiovascular monitoring research and the development of personal health applications.

Prediction of carbon emissions from municipal solid waste treatment in 31 provinces and cities in China under the shared socio-economic pathways

In order to reduce greenhouse gas (GHG) emissions from the municipal solid waste (MSW) sector, it is essential to describe its greenhouse gas emission patterns and propose appropriate mitigation measures. Therefore, this study predicts the greenhouse gas emissions of urban domestic waste treatment by combining the Intergovernmental Panel on Climate Change (IPCC), shared socio-economic pathways and in-depth learning model. Under the hypothetical policy scenario and the shared socio-economic pathways (SSPs), the greenhouse gas emissions from municipal solid waste treatment in 31 provinces and cities in China by 2030 are analyzed. Combined with the model comparison results, the bidirectional long and short-term memory neural network model has better prediction accuracy than the other models, with an average mean absolute percentage error of 9.68% for 31 provinces and cities, indicating good applicability. Scenario analysis shows that greenhouse gas emissions from municipal solid waste disposal can be significantly reduced only after the implementation of municipal solid waste classification. Scenario 3 (S3) of waste segregation will produce the most GHG emissions, and switching to Scenario 5 (S5) of waste segregation will produce significantly less GHG emissions. Finally, the paper makes policy recommendations on the priority areas of urban domestic waste classification, waste source classification and recycling, and the structure of urban municipal solid waste treatment.

Continuous Assessment of Function and Disability via Mobile Sensing: Real-World Data-Driven Feasibility Study.

Functional limitations are associated with poor clinical outcomes, higher mortality, and disability rates, especially in older adults. Continuous assessment of patients' functionality is important for clinical practice; however, traditional questionnaire-based assessment methods are very time-consuming and infrequently used. Mobile sensing offers a great range of sources that can assess function and disability daily. This work aims to prove the feasibility of an interpretable machine learning pipeline for predicting function and disability based on the World Health Organization Disability Assessment Schedule (WHODAS) 2.0 outcomes of clinical outpatients, using passively collected digital biomarkers. One-month-long behavioral time-series data consisting of physical and digital activity descriptor variables were summarized using statistical measures (minimum, maximum, mean, median, SD, and IQR), creating 64 features that were used for prediction. We then applied a sequential feature selection to each WHODAS 2.0 domain (cognition, mobility, self-care, getting along, life activities, and participation) in order to find the most descriptive features for each domain. Finally, we predicted the WHODAS 2.0 functional domain scores using linear regression using the best feature subsets. We reported the mean absolute errors and the mean absolute percentage errors over 4 folds as goodness-of-fit statistics to evaluate the model and allow for between-domain performance comparison. Our machine learning-based models for predicting patients' WHODAS functionality scores per domain achieved an average (across the 6 domains) mean absolute percentage error of 19.5%, varying between 14.86% (self-care domain) and 27.21% (life activities domain). We found that 5-19 features were sufficient for each domain, and the most relevant being the distance traveled, time spent at home, time spent walking, exercise time, and vehicle time. Our findings show the feasibility of using machine learning-based methods to assess functional health solely from passively sensed mobile data. The feature selection step provides a set of interpretable features for each domain, ensuring better explainability to the models' decisions-an important aspect in clinical practice.