Performance Evaluation Of Algorithms Research Articles

Advancements in Optimization: Critical Analysis of Evolutionary, Swarm, and Behavior-Based Algorithms

The research work on optimization has witnessed significant growth in the past few years, particularly within multi- and single-objective optimization algorithm areas. This study provides a comprehensive overview and critical evaluation of a wide range of optimization algorithms from conventional methods to innovative metaheuristic techniques. The methods used for analysis include bibliometric analysis, keyword analysis, and content analysis, focusing on studies from the period 2000–2023. Databases such as IEEE Xplore, SpringerLink, and ScienceDirect were extensively utilized. Our analysis reveals that while traditional algorithms like evolutionary optimization (EO) and particle swarm optimization (PSO) remain popular, newer methods like the fitness-dependent optimizer (FDO) and learner performance-based behavior (LPBB) are gaining attraction due to their adaptability and efficiency. The main conclusion emphasizes the importance of algorithmic diversity, benchmarking standards, and performance evaluation metrics, highlighting future research paths including the exploration of hybrid algorithms, use of domain-specific knowledge, and addressing scalability issues in multi-objective optimization.

A novel algorithm combined X-ray fluorescence and Neural Network (XRF-NN) for coal ash content prediction: Algorithm design and performance evaluation

This paper investigated a precise algorithm combining X-ray Fluorescence and Neural Network (XRF-NN) for predicting ash content. The 261 sets of XRF tests show that the 34 elements in the chosen Guqiao coal can be categorized as major, secondary and tiny elements, whose cumulative ratio were ~95%, 4-5% and <1%, respectively. Referring to the machine learning theory, the construction strategy of Element-Ash dataset was determined viz. value determination → standardization → division → optimization of generalization ability. Then, the hyperparameters optimization displays that XRF-NN model with 34 inputs and 1 output was suitable to predict coal ash content, where the activation function, loss function, and optimizer were ReLU, MAE, and Momentum, respectively. After iterative training, the new XRF-NN model provides the precise prediction of coal ash content with the absolute errors between -2.0% and 2%. Moreover, the prediction accuracy rose from 57.69% to 100%, as the expected relative error increased from 1% to 5%. Furthermore, the comparisons between different prediction methods reveal that the minimum MRE of 1.22% can be obtained by XRF-NN with total elements, which was only half of those given by the conventional Multiple Linear Regression and Partial Least Squares Regression. Besides, XRF-NN model presents the root mean squared error 0.797%, the mean absolute error 0.625% and the coefficient of determination 0.999, which were significantly superior to those calculated by Dual-Energy Gamma-ray Transmission, Ploy, RFR, XGBoost and DNN model. The results of this study suggest the excellent performance of the new XRF-NN model in predicting ash content.

Performance Evaluation of Machine Learning Algorithms for Iris Dataset Classification

The problem of classifying iris flowers is a well-known issue in the fields of machine learning and pattern recognition. The Iris dataset is widely recognized as a standard for evaluating the effectiveness of classification algorithms in machine learning. This paper offers a comparative assessment of different machine learning algorithms when applied to the Iris dataset. We implement and assess the performance of k-Nearest Neighbors, Logistic Regression, Decision Tree, Linear SVC methods, Random Forest, Gaussian Naïve Bayes and AdaBoost algorithms based on their accuracy in classification. We examine diverse classification methods, appraise their effectiveness, and discuss our findings. The outcomes reveal the accuracy of these models in correctly categorizing the Iris species. To achieve high accuracy, k-Nearest Neighbors, Logistic Regression, Linear SVC methods,Decision Tree,, Random Forest and AdaBoost classifiers are utilised.

Performance Evaluation of Machine Learning Algorithms in Reduced Dimensional Spaces

Performance Evaluation of Machine Learning Algorithms in Reduced Dimensional Spaces

A review of the development of YOLO object detection algorithm

The You Only Look Once (YOLO) algorithm series, as the forefront of object detection technology, has evolved from YOLOv1 to YOLOv10, consistently enhancing detection speed and accuracy. Through literature review and data analysis. This paper mainly discusses the development processes of the YOLO algorithm series, focuses on the changes and innovations in network structure, training strategies, and performance optimization. By introducing techniques, such as CSPNet, Anchor-free, data augmentation, and multi-scale training, the YOLO algorithm has progressively found a better balance between detection speed and accuracy, demonstrating excellent performance in processing real-time images and handling high-complexity scenarios. Furthermore, this paper also addresses some challenges faced by the YOLO algorithms and potential future research directions, such as the lower accuracy in detecting small targets and reduced robustness in complex scenarios. Through analysis, potential optimization directions for the future include further refining network structure and employing more efficient training methods to enhance algorithm efficiency. This paper intends to offer a thorough performance evaluation of the YOLO series algorithms, identify potential areas for future improvements to advance YOLO technology.

Early classification and recognition algorithm for sudden cardiac arrest based on limited electrocardiogram data trained with a two-stages convolutional neural network

Sudden cardiac arrest (SCA) is a lethal cardiac arrhythmia that poses a serious threat to human life and health. However, clinical records of sudden cardiac death (SCD) electrocardiogram (ECG) data are extremely limited. This paper proposes an early prediction and classification algorithm for SCA based on deep transfer learning. With limited ECG data, it extracts heart rate variability features before the onset of SCA and utilizes a lightweight convolutional neural network model for pre-training and fine-tuning in two stages of deep transfer learning. This achieves early classification, recognition and prediction of high-risk ECG signals for SCA by neural network models. Based on 16 788 30-second heart rate feature segments from 20 SCA patients and 18 sinus rhythm patients in the international publicly available ECG database, the algorithm performance evaluation through ten-fold cross-validation shows that the average accuracy (Acc), sensitivity (Sen), and specificity (Spe) for predicting the onset of SCA in the 30 minutes prior to the event are 91.79%, 87.00%, and 96.63%, respectively. The average estimation accuracy for different patients reaches 96.58%. Compared to traditional machine learning algorithms reported in existing literatures, the method proposed in this paper helps address the requirement of large training datasets for deep learning models and enables early and accurate detection and identification of high-risk ECG signs before the onset of SCA.

Performance and energy evaluation of dynamic adaptive deterministic routing algorithm for multicore architectures

Chip Multiprocessors (CMPs) leverage multiple processing units to improve computational speed and efficiency. Routing algorithms in NoC (Network-on-Chip) architectures ensure efficient data communication between these units, addressing challenges such as congestion, latency, and power consumption. This paper delves into an extensive analysis of the performance evaluation concerning the Dynamic Adaptive Deterministic (DyAD) routing for Network-on-Chip (NoC) and Wireless Network-on-Chip (WiNoC) configurations within the framework of a 64-core and 100-core multicore architecture. The evaluation involved a congestion threshold analysis of data transmission latency, the efficiency of network data throughput, and the amount of energy consumed. To substantiate our conclusions, we conducted simulations of the 64-core and 100-core mesh-based NoC and WiNoC architectures under the Video Image Processing System (VIPS) benchmark workload. These simulations were executed utilizing the Noxim simulator, a well-recognized tool acclaimed for its capacity to provide cycle-accurate simulations. Analyzing the simulation outcomes, it becomes evident that the DyAD routing approach for the 64-core and 100-core WiNoC architecture performs better in terms of network performance. It is characterized by enhanced capabilities at higher Packet Injection Rate (PIR) saturation loads, improved network throughput, and minimized latencies. This dominance is evident from its ability to handle heavier workloads and achieve lower delays in the investigated VIPS workload situations, compared to the 64-core and 100-core NoC multicore architecture.

Classification of Firewall Log Files with Different Algorithms and Performance Analysis of These Algorithms

Classifying firewall log files allows analysing potential threats and deciding on appropriate rules to prevent them. Therefore, in this study, firewall log files are classified using different classification algorithms and the performance of the algorithms are evaluated using performance metrics. The dataset was prepared using the log files of a firewall. It was filtered to make it free from any personal data and consisted of 12 attributes in total and from these attributes the action attribute was selected as the class. In the performance evaluation, Simple Cart and NB tree algorithms made the best predictions, achieving an accuracy rate of 99.84%. Decision Stump had the worst prediction performance, achieving an accuracy rate of 79.68%. As the total number of instances belonging to each of the classes in the dataset was not equal, the Matthews correlation coefficient was also used as a performance metric in the evaluations. The Simple Cart, BF tree, FT tree, J48 and NB Tree algorithms achieved the highest average values. However, although the reset-both class was not predicted successfully by the others, the Simple Cart algorithm made the best predictions for it. The values of other performance metrics used in this study also support this conclusion. Therefore, the Simple Cart algorithm is recommended for use in classifying firewall log files. However, there is a need to develop a prefiltering and parsing approach to process different log files as each firewall brand creates and maintains log files in its own format. Therefore, in this study, a novel prefiltering and parsing approach has been proposed to process log files with different structures and create structured datasets using them.

Construction of Risk Prediction Models for Enterprise Finance Sharing Operations Using K-Means and C4.5 Algorithms

The evaluation of financial sharing centres in enterprises typically relies on outdated financial data, lacks comprehensive assessment, and presents risks such as employee misconduct. To address these challenges, we propose a risk prediction model for enterprise financial sharing operations based on the K-means clustering algorithm for performance evaluation and the C4.5 algorithm for managing employee risks. Our approach enhances the accuracy and objectivity of performance evaluation while improving the efficiency of personnel risk management. Results indicate that the K-means algorithm classifies employee performance into five levels, facilitating comprehensive performance evaluation. Furthermore, through risk management optimisation, accuracy and recall rates increase to 0.905 and 0.890, respectively. The proposed risk prediction model achieves high accuracy rates of 90.5% and 92.4% in the training and test sets, respectively. Practical application of our methodology and model in A Group's financial sharing centre demonstrates their effectiveness and potential for enhancing the operation and management of enterprise financial sharing centres.

Scalable and Distributed Mathematical Modeling Algorithm Design and Performance Evaluation in Heterogeneous Computing Clusters

Scalable and Distributed Mathematical Modeling Algorithm Design and Performance Evaluation in Heterogeneous Computing Clusters

Performance Evaluation of Machine Learning Algorithms in Smart Agriculture

Performance Evaluation of Machine Learning Algorithms in Smart Agriculture

Performance evaluation of MPPT algorithm of VIPV systems in realistic urban routes using image processing

The assessment of the power that Vehicle Integrated Photovoltaic (VIPV) systems produce during the vehicle displacement in urban environments requires to know the effect of the dynamic shadowing over the photovoltaic (PV) surface and the response that the DC/DC electronics have over these situations of fast-evolving shadows. This paper poses a novel methodology that evaluate the shadowing profile on the vehicle through images and estimates the PV output power and the behavior of a classical Maximum Power Point Tracking (MPPT) algorithm through simulations. Through this methodology, the efficiency (intended as the ratio between the maximum power generated by the PV module and the power delivered by the MPPT algorithm) has been evaluated in a real dynamic shadowing situation, where images of the vehicle roof in a 22 km route have been acquired at high frequency.

Retraction Note: Database task processing optimization based on performance evaluation and machine learning algorithm

Retraction Note: Database task processing optimization based on performance evaluation and machine learning algorithm

Performance evaluation of path choice set generation algorithms for route choice modelling

Traffic assignment and its applications are sensitive to the selection of path choice sets between an OD pair. In literature, different path generation algorithms are proposed, namely, Link Labeling, Link Elimination, Link Penalty, Simulation, and Branch and Bound. The effectiveness of these algorithms hinges on algorithmic principles and input parameters. Consequently, choosing an appropriate algorithm for practical implementation becomes challenging. This paper aims to benchmark these algorithms with a case study using a real vehicle trajectory dataset from four OD pairs from Brisbane, Australia. The paper offers both qualitative and quantitative comparisons of the algorithms. The performance of the algorithms is evaluated based on their authenticity, redundancy, and applicability. The results highlight that the simulation method outperforms the other algorithms, specifically utilizing a truncated normal distribution. The paper offers valuable insights into the selection of path-generation techniques, aiding the enhancement of traffic assignment processes.

Optimization research on energy consumption and carbon emissions in hydrogen-blended natural gas pipeline networks considering compressor performance - A case study of the west-east gas pipeline (line 1)

The blending of hydrogen into natural gas can utilize existing natural gas pipeline infrastructure to achieve long-distance, large-scale, and cost-effective hydrogen transportation. However, the introduction of hydrogen poses a series of new technical and safety challenges to the operation of natural gas pipelines. Addressing these issues, this paper focuses on minimizing compressor station energy consumption and carbon emissions as objective functions, conducting a comprehensive analysis of station operation modes. Incorporating the characteristics of the model, the soft frost search strategy, the hard frost penetration mechanism, and the positive greedy selection mechanism are introduced to propose the IRIME algorithm. Comparisons are made with well-performing optimization algorithms using classical test functions to evaluate algorithm performance. To further validate the algorithm's accuracy, the West-East Gas Pipeline (Line 1) is considered, comparing the convergence speed and precision of the IRIME algorithm, leading to its selection for solving the model. Furthermore, consider analyzing the impact of different hydrogen blending ratios on the performance curves and operating points of compressors. Results indicate that compared to conditions without hydrogen blending, the compressor outlet temperature and pressure drop increase for hydrogen blending ratios of 5%, 10%, 15%, and 20%; as the hydrogen blending ratio increases, the efficiency of fuel-driven compressors, efficiency of electrically driven compressors, and average efficiency of compressor units decrease; with an increase in hydrogen blending ratio, carbon emissions from fuel-driven compressors and electric-driven compressors both decrease. In conclusion, with an increasing hydrogen blending ratio, total energy consumption increases by 3.14%, 5.69%, 8.68%, and 10.90%, while total carbon emissions decrease by 2.99%, 5.36%, 8.32%, and 10.68%. These findings provide valuable insights for operational guidance in the field.

Automated detection of landslide events from multi-source remote sensing imagery: Performance evaluation and analysis of YOLO algorithms

Automated detection of landslide events from multi-source remote sensing imagery: Performance evaluation and analysis of YOLO algorithms

Comparative Analysis of Machine Learning Algorithm Performance in Predicting Stunting in Toddlers

Stunting is a condition where the growth of children and toddlers is stunted, which causes children to be shorter than they should be. In the long term, stunting can reduce reproductive health, study concentration, and work productivity, thereby causing significant state losses. The prevalence of stunting in Indonesia, which is still above 20 percent, shows that there are still chronic nutritional problems among toddlers. To prevent this from happening, identification as early as possible can be done using machine learning for predictions. The aim of this research is to conduct a comparative analysis of the performance of machine learning algorithms for predicting stunting in toddlers. Random Forest, K-Nearest Neighbors, and Extreme Gradient Boosting are the algorithms that are compared for their performance. The performance of each algorithm is measured using evaluation matrices such as accuracy, precision, recall, and f1-score. The research method starts with data collection, data preprocessing, data splitting, application of machine learning algorithms, evaluation of algorithm performance, and comparison of results. The performance evaluation matrix measurement results show that Random Forest has an accuracy of 99.95%, precision of 99.89%, recall of 99.94%, and f1-score of 99.91%. K-Nearest Neighbors has an accuracy of 99.93%, precision of 99.87%, recall of 99.88%, and f1-score of 99.88%. Meanwhile, Extreme Gradient Boosting has an accuracy of 99.36%, precision of 98.86%, recall of 98.95%, and f1-score of 98.90%. From the results of all performance evaluation matrices, it can be concluded that the random forest algorithm is the best algorithm for predicting stunting in toddlers.

Detection and Analysis of Corrosion on Coated Metal Surfaces Using Enhanced YOLO v5 Algorithm for Anti-Corrosion Performance Evaluation

This study addresses the severe corrosion issues in the coastal regions of southern China by proposing an improved YOLO v5-GOLD-NWD model. Utilizing corrosion data from the National Center for Materials Corrosion and Protection Science of China, a dataset was constructed for metal-surface corrosion under different protective coatings. This dataset was used for model training, testing, and comparison. Model accuracy was validated using precision, recall, F1 score, and prediction probability. The results demonstrate that the proposed improved model exhibits better identification precision in metal corrosion detection, achieving 78%, a 4% improvement compared to traditional YOLO v5 models. Additionally, through identification and statistical analysis of corrosion image datasets from five types of coated metal specimens, it was found that powder epoxy coating, fluorocarbon coating, epoxy coating, and chlorinated rubber coating showed good corrosion resistance after 24 months of exposure. Conversely, Wuxi anti-fouling coating exhibited poor corrosion resistance. After 60 months of natural exposure, the powder epoxy coating specimens had the highest corrosion occurrence probability, followed by chlorinated rubber coating and epoxy coating, with fluorocarbon coating showing relatively lower probability. The fluorocarbon coating demonstrated relatively good corrosion resistance at both 24 and 60 months of exposure. The findings of this study provide a theoretical basis for enhancing the corrosion protection effectiveness of steel structures in coastal areas.

An Agile Approach for Adopting Sustainable Energy Solutions with Advanced Computational Techniques

In the face of the burgeoning electricity demands and the imperative for sustainable development amidst rapid industrialization, this study introduces a dynamic and adaptable framework suitable for policymakers and renewable energy experts working on integrating and optimizing renewable energy solutions. While using a case study representative model for Sub-Saharan Africa (SSA) to demonstrate the challenges and opportunities present in introducing optimization methods to bridge power supply deficits and the scalability of the model to other regions, this study presents an agile multi-criteria decision tool that pivots on four key development phases, advancing established methodologies and pioneering refined computational techniques, to select optimal configurations from a set of Policy Decision-Making Metrics (PDM-DPS). Central to this investigation lies a rigorous comparative analysis of variants of three advanced algorithmic approaches: Swarm-Based Multi-objective Particle Swarm Optimization (MOPSO), Decomposition-Based Multi-objective Evolutionary Algorithm (MOEA/D), and Evolutionary-Based Strength Pareto Evolutionary Algorithm (SPEA2). These are applied to a grid-connected hybrid system, evaluated through a comprehensive 8760-hour simulation over a 20-year planning horizon. The evaluation is further enhanced by a set of refined Algorithm Performance Evaluation Metrics (AL-PEM) tailored to the specific constraints. The findings not only underscore the robustness and consistency of the SPEA2 variant over 15 runs of 200 generations each, which ranks first on the AL-PEM scale, but the findings also validate the strategic merit of combining multiple technologies and empowering policymakers with a versatile toolkit for informed decision-making.

A Time-Domain Doppler Estimation and Waveform Recovery Approach with Iterative and Ensemble Techniques for Bi-Phase Code in Radar Systems

This paper presents a novel, cost-effective technique for estimating the Doppler effect in the time domain using a single pulse and subsequently leveraging the precise Doppler value to recover the radar waveform. The proposed system offers several key advantages over existing techniques, including the ability to calculate the target speed without any frequency ambiguity and the ability to detect a wide range of target speeds. These two features are not available in any existing techniques, including the conventional moving target detection (MTD) processor. To ensure improved accuracy and robust estimation, the system employs ensemble and iterative techniques by recursively and efficiently reducing the Doppler residues from the signal. Furthermore, the proposed system demonstrates effective signal recovery of a well-known bi-phase code shape at low signal-to-noise ratios in just a few iterations. The performance evaluation of the new algorithm demonstrates its practicability and its superiority over traditional radar systems. Implementation on software-defined radio (SDR) reveals that the proposed system excels in Doppler estimation and signal recovery at low SNRs, demonstrating promising results.