Performance Of Algorithm Research Articles

Advancements in evaporation prediction: introducing the Gated Recurrent Unit–Multi-Kernel Extreme Learning Machine (MKELM)–Gaussian Process Regression (GPR) model

Predicting evaporation is an essential topic in water resources management. It is critical to plan irrigation schedules, optimize hydropower production, and accurately calculate the overall water balance. Thus, researchers have developed many prediction models for predicting evaporation. Despite the development of these models, there are still unresolved challenges. These challenges include selecting the most important input parameters, handling nonstationary data, extracting critical information from data, and quantifying the uncertainty of predicted values. Thus, the main aim of this study is to address these challenges by developing a new prediction model. The new prediction model, named Gated Recurrent Unit–Multi-Kernel Extreme Learning Machine (MKELM)–Gaussian Process Regression (GPR), was used to predict one-month ahead evaporation in the Kashafrood basin, Iran. This model was executed in multiple stages. First, a feature selection algorithm was used to determine the most critical input parameters. A data processing technique was then employed to decompose nonstationary data into stationary intrinsic mode functions (IMFs). The GRU model then processed these components to extract their essential information. In the following step, the extracted information was inserted into the MKELM model to predict evaporation. Finally, the GPR model quantified the uncertainty of predicted values. Our research also introduces a new optimizer called the Salp Swarm Optimization Algorithm–Sine Cosine Optimization Algorithm. This algorithm was used to tune the model parameters. This algorithm's performance and the prediction models’ accuracy were evaluated using several error indices. According to the study results, the GRU–MKELM–GPR model performed better than other models in predicting monthly evaporation. It improved the training and testing mean absolute error values of the other models by 21%-43% and 8.2–33%, respectively. Moreover, the new model improved the R2 (R-squared or coefficient of determination) values of other models by 5–12%. Generally, the main findings of this paper included the superior performance of the new model in predicting evaporation data and the superior performance of a new optimizer in adjusting model parameters. These findings highlighted the effectiveness of the suggested model in addressing the challenges associated with evaporation prediction.

Dual-weight decay mechanism and Nelder-Mead simplex boosted RIME algorithm for optimal power flow

The increasing demand for electricity presents substantial challenges in power system planning, particularly optimizing the Optimal Power Flow (OPF) problem. The OPF problem entails establishing the best settings for control variables in a power system to reduce objectives such as generating cost and transmission losses while meeting operational restrictions. This research introduces an upgraded RIME optimization algorithm (WDNMRIME) to address these challenges. WDNMRIME integrates a dual-weight decay mechanism and the Nelder-Mead simplex (NMs), enhancing population diversity and mitigating the risk of local optima. Additionally, NMs expedites convergence by refining the population's optimal solution set. Experimental validation on the IEEE 30-bus test system demonstrates that WDNMRIME achieves a generation cost of $806.00298 per hour and reduces total power loss from 1.43 MW to 1.39 MW. These results surpass the performance of the original RIME algorithm, showcasing a 15% improvement in convergence speed. The algorithm effectively optimizes multiple concurrent Flexible Alternating Current Transmission Systems (FACTS) devices, even under the uncertain nature of wind energy resources modeled using the Weibull probability density function. These findings highlight WDNMRIME's significant contribution to improving OPF optimization in dynamic power systems.

Wireless UV collaborative RSSI and the AOA hybrid localization method for UAV swarms

Wireless ultraviolet (UV) light can guarantee the inter-copter positioning accuracy of cluster-flying unmanned aerial vehicles (UAVs) in an electromagnetic confrontation environment. By combining wireless UV received signal strength indication (RSSI) localization and the angle of arrival (AOA) localization algorithm, a UV-hybrid localization method is proposed for UV communication collaboration between UAV swarms. The method collects the UV signal strength between the anchor UAV node and the unknown UAV node to obtain the inter-aircraft distance information, establishes a Gaussian hybrid noise model based on semi-definite relaxation, and uses the UV angle of the arrival estimation to solve the angle between the UAVs to achieve the maximum likelihood estimation of node positions in UAV swarms. A simulation comparison of the UV-hybrid localization algorithm, weighted least squares, and the Gaussian hybrid semi-definite planning localization algorithm is carried out. The results show that the performance of the UV-hybrid localization algorithm is close to the Cramer–Rao lower bound, and the average localization error is reduced by 32.9% and 15.6% compared to weighted least squares and Gaussian hybrid semi-definite planning algorithms; the algorithm of this paper achieves the node localization with fewer iterations, and it has a higher accuracy and efficiency of localization than the other algorithms.

Selecting the Right Metric: A Detailed Study on Image Segmentation Evaluation

<p dir="ltr"><span>Image segmentation is critical role in various computer vision tasks, such as object recognition, medical imaging, autonomous driving, and document analysis. To evaluate the performance of segmentation algorithms, various metrics have been developed, each providing insights into different aspects of segmentation accuracy, object completeness, and boundary precision. This paper presents a comprehensive review of the key evaluation metrics used in image segmentation, including popular metrics such as Intersection over Union, Dice Coefficient, F1-score, and boundary-based metrics like average and maximum boundary distance. We explore the taxonomy of evaluation techniques, distinguishing between supervised and unsupervised methods, and discuss the strengths, limitations, and applications of each metric. Furthermore, we address the challenges associated with image segmentation, including handling noise, occlusions, illumination variation, and the scarcity of annotated datasets for training. By offering a detailed analysis of evaluation metrics, this paper aims to guide researchers and practitioners in selecting appropriate metrics for specific tasks and improving algorithm performance through informed comparisons and parameter tuning.</span></p><div><span><br /></span></div>

UAV Swarm Rounding Strategy Based on Deep Reinforcement Learning Goal Consistency with Multi-Head Soft Attention Algorithm

Aiming at the problem of target rounding by UAV swarms in complex environments, this paper proposes a goal consistency reinforcement learning approach based on multi-head soft attention (GCMSA). Firstly, in order to make the model closer to reality, the reward function when the target is at different positions and the target escape strategy are set, respectively. Then, the Multi-head soft attention module is used to promote the information cognition of the target among the UAVs, so that the UAVs can complete the target roundup more smoothly. Finally, in the training phase, this paper introduces cognitive dissonance loss to improve the sample utilization. Simulation experiments show that GCMSA is able to obtain a higher task success rate and is significantly better than MADDPG in terms of algorithm performance.

EKFNet: Edge-based Kalman filter network for real-time EEG signal denoising.

Signal denoising methods based on deep learning have been extensively adopted on Electroencephalogram(EEG) devices. However, they are unable to deploy on edge-based portable or wearable (P/W) electronics due to the high computational complexity of the existed models. To overcome such issue, we propose an edge-based lightweight Kalman filter network (EKFNet) that does not require manual prior knowledge estimation.
Approach: Specifically, we construct a multi-scale feature fusion (MSFF) module to capture multi-scale feature information and implicitly compute the prior knowledge. Meanwhile, we design an adaptive gain estimation (AGE) module that incorporates long short-term memory (LSTM) and sequential channel attention module (CAM) to dynamically predict the Kalman gain. Furthermore, we present an optimization strategy utilizing operator fusion and constant folding to reduce the model's computational overhead and memory footprint.
Main results: Experimental results show that the EKFNet reduces the sum of the square of the distances by at least 12% and improves the cosine similarity by at least 2.2% over the state-of-the-art methods. Besides, the model optimization shortens the inference time by approximately 3.3×. The code of our EKFNet is available at https://github.com/cathnat/EKFNet.
Significance: By integrating Kalman filter with deep learning, the approach addresses the parameter-setting challenges in traditional algorithms while reducing computational overhead and memory consumption, which exhibits a good tradeoff between algorithm performance and computing power.

Fault diagnosis of belt conveyors using audio data based on LLE with adaptive neighborhood and neighbor optimization under multi-information fusion

To solve the problem of insufficient dimensionality reduction performance of dimensionality reduction algorithms in fault diagnosis of belt conveyors using audio data (FDBCA), we propose a locally linear embedding (LLE) algorithm with adaptive neighborhood and neighbor optimization under multi-information fusion (FM-ANO-LLE) for FDBCA. This method first uses a multi-information fusion metric to assess the relationships between samples from multiple perspectives and improve the accuracy of the selected samples. Second, an adaptive neighborhood strategy based on sample metric discriminatory information is used to dynamically divide the local structure of the sample. Finally, the reconstruction weights are optimized through the nearest neighbor optimization strategy to improve the dimensionality reduction performance. The experimental results demonstrated that FM-ANO-LLE outperforms existing dimensionality reduction algorithms, the performance of the dimensionality reduction algorithm is significantly improved, and it achieves higher diagnostic accuracy in practical FDBCA, with an accuracy of 90.4%.

Property Valuation: Using Machine Learning for Rent Value Prediction

The rising demand for rental spaces, such as hotel lodging rooms, has led property owners in high-potential areas to begin renting out their properties to meet this need. The Directorate General of State Assets (DJKN) under the Ministry of Finance, which manages state-owned assets, oversees approximately 1,100 room units with potential for rental. In alignment on increasing non-tax state revenue from state-owned assets (BMN), offering lodging room rentals as a form of BMN utilization presents a promising opportunity for expansion. To facilitate efficient rental services for these BMN rooms, a quick and accurate valuation process is essential. However, establishing fair rental prices for both property owners and tenants poses a significant challenge. This study, therefore, aims to develop automated valuation model (AVM). Since rental price predictions depend on various bebas variabels, the study employs several machine learning algorithms, including Ridge, Linear Regression, Random Forest (RF), Lasso, SVR, Elastic Net, and Extreme Gradient Boosting. The research focuses on rental markets in Jakarta, Bogor, and Bandung, using a dataset of fewer than 500 rental room samples sourced from marketplaces. These data points are adjusted for state-owned asset variabels, followed by feature engineering. The performance of the various algorithms is then compared. The results indicate that the best-performing model and the most important features vary by city, but overall, the Random Forest model delivers strong performance, with room size, road width, and room quality emerging as the most influential factors.

Machine learning complete intersection Calabi-Yau 3-folds

Gaussian process regression, kernel support vector regression, the random forest, extreme gradient boosting, and the generalized linear model algorithms are applied to data of complete intersection Calabi-Yau threefolds. It is shown that Gaussian process regression is the most suitable for learning the Hodge number h2,1 in terms of h1,1. The performance of this regression algorithm is such that the Pearson correlation coefficient for the validation set is R2=0.9999999995 with a root mean square error RMSE=0.0002895011. As for the train set, these two parameters are as follows: R2=0.9999999994 and RMSE=0.0002854348. The training error and the cross-validation error of this regression are 1×10−9 and 1.28×10−7, respectively. Learning the Hodge number h1,1 in terms of h2,1 yields R2=1.000000 and RMSE=7.395731×10−5 for the validation set of the Gaussian process regression. Published by the American Physical Society 2024

Hybrid Population-Based Hill Climbing Algorithm for Generating Highly Nonlinear S-boxes

This paper introduces the hybrid population-based hill-climbing (HPHC) algorithm, a novel approach for generating cryptographically strong S-boxes that combines the efficiency of hill climbing with the exploration capabilities of population-based methods. The algorithm achieves consistent generation of 8-bit S-boxes with a nonlinearity of 104, a critical threshold for cryptographic applications. Our approach demonstrates remarkable efficiency, requiring only 49,277 evaluations on average to generate such S-boxes, representing a 600-fold improvement over traditional simulated annealing methods and a 15-fold improvement over recent genetic algorithm variants. We present comprehensive experimental results from extensive parameter space exploration, revealing that minimal populations (often single-individual) combined with moderate mutation rates achieve optimal performance. This paper provides detailed analysis of algorithm behavior, parameter sensitivity, and performance characteristics, supported by rigorous statistical evaluation. We demonstrate that population size should approximate available thread count for optimal parallel execution despite smaller populations being theoretically more efficient. The HPHC algorithm maintains high reliability across diverse parameter settings while requiring minimal computational resources, making it particularly suitable for practical cryptographic applications.

An Evaluation of an Algorithm for the Selection of Flexible Survival Models for Cancer Immunotherapies: Pass or Fail?

Accurately extrapolating survival beyond trial follow-up is essential in a health technology assessment where model choice often substantially impacts estimates of clinical and cost effectiveness. Evidence suggests standard parametric models often provide poor fits to long-term data from immuno-oncology trials. Palmer et al. developed an algorithm to aid the selection of more flexible survival models for these interventions. We assess the usability of the algorithm, identify areas for improvement and evaluate whether it effectively identifies models capable of accurate extrapolation. We applied the Palmer algorithm to the CheckMate-649 trial, which investigated nivolumab plus chemotherapy versus chemotherapy alone in patients with gastroesophageal adenocarcinoma. We evaluated the algorithm's performance by comparing survival estimates from identified models using the 12-month data cut to survival observed in the 48-month data cut. The Palmer algorithm offers a systematic procedure for model selection, encouraging detailed analyses and ensuring that crucial stages in the selection process are not overlooked. In our study, a range of models were identified as potentially appropriate for extrapolating survival, but only flexible parametric non-mixture cure models provided extrapolations that were plausible and accurately predicted subsequently observed survival. The algorithm could be improved with minor additions around the specification of hazard plots and setting out plausibility criteria. The Palmer algorithm provides a systematic framework for identifying suitable survival models, and for defining plausibility criteria for extrapolation validity. Using the algorithm ensures that model selection is based on explicit justification and evidence, which could reduce discordance in health technology appraisals.

A Hybrid Optimization Algorithm Based on Cascaded (1+PI)-PI-PID Controller for Load Frequency Control in Interconnected Power Systems

Electric power generation is crucial for electrical systems, and load frequency control (LFC) ensures system synchronization and reliability. Modern systems use well-designed controllers, with selecting the right type and design approach being essential for a good LFC controller design. This study introduces a cascaded (1+PI)-PI-PID controller optimized using the improved Artificial Bee Colony-Particle Swarm Optimization (IABC-PSO) algorithm for LFC in interconnected power systems, aiming to address complex challenges associated with LFC in modern power systems. The research examines a two-area power system model with hydro and thermal power plants and a capacitive energy storage (CES) device. It proposes a (1+PI)-PI-PID controller structure that integrates traditional PID controllers, enhancing performance and debugging capabilities. The study uses integral absolute error (IAE), integral square error (ISE), integral time absolute error (ITAE), and integral time square error (ITSE) as objective functions. The IABC-PSO algorithm introduces a novel decision block to increase scout bee (SB) occurrences to 10% for each employed bee (EB), and a new control search phase to optimize SBs' exploration capabilities in a controlled manner, enhancing their efficiency. The algorithm's performance was assessed using four benchmark functions, revealing superior results compared to the Hybrid ABC-PSO algorithm. The study examined six experimental configurations of the controller, varying the sequence of (1+PI), PI, and PID components. The proposed strategy was tested under 1% and 5% step load perturbations in area-1 and area-2 to evaluate dynamic performance. This study conducts a sensitivity analysis by adjusting hydro parameters within a ±25% range, and considers generation rate constraint (GRC) and boiler dynamics (BD) in a two-area reheat thermal-hydro-power system model for reliability and credibility. The proposed IABC-PSO algorithm outperforms traditional methods (ABC, PSO, and hybrid ABC-PSO, WOAw, GWO, hSFS-LUS, and FA) in optimizing parameters, achieving shortest settling times and minimal ITAE values. Its performance is evaluated under random step load perturbations (SLP) and its internal robustness is demonstrated through stability analysis using the Lyapunov method. In conclusion, the hybrid IABC-PSO algorithm-based cascaded (1+PI)-PI-PID controller offers a robust and efficient solution for LFC in interconnected power systems. It ensures enhanced performance, stability, and adaptability, making it a valuable contribution to modern power system management.

Many-Objective Jaccard-Based Evolutionary Feature Selection for High-Dimensional Imbalanced Data Classification.

Filters and wrappers represent two mainstream approaches to feature selection (FS). Although evolutionary wrapper-based FS outperforms filters in addressing real-world classification problems, extending these methods to high-dimensional, many-objective optimization problems with imbalanced data poses substantial challenges. Overcoming computational costs and identifying suitable performance metrics are vital for navigating search operation complexities. Here, we propose using the Jaccard similarity (JS) in a set-based evolutionary many-objective (JSEMO) FS search, addressing both evolutionary FS and imbalanced classifier choice concurrently. This study highlights the mutual influence between these aspects, impacting overall algorithm performance. JSEMO integrates JS into population initialization, reproduction, and elitism steps, enhancing diversity and avoiding duplicate solutions. The set-based variation operator utilizes intersection and union operators for compatibility with binary coding. We also introduce a double-weighted KNN (KNN2W) classifier with four supportive objectives as a many-objective FS problem to handle imbalanced distributions. Compared with 20 methods on 15 benchmark problems, JSEMO produces distinct optimal features, significantly improving overall accuracy, balance accuracy, and g-mean metrics with comparable feature set size and computational cost. The ablation study underscores the positive impact of all JSEMO components, highlighting the set-based variation operation with JS and KNN2W with relevant evaluation metrics as the most influential aspects.

A decomposition-based multi-objective evolutionary algorithm using infinitesimal method

Multi-Objective Evolutionary Algorithm based on decomposition (MOEA/D) has been extensively employed to address a diverse array of real-world challenges and has shown excellent performance. However, the initial collection of aggregate weight vectors proves unsuitable for multi-objective optimization problems (MOPs) featuring intricate Pareto front (PF) structures, and the solving performance will be greatly affected when MOEA/D solves these irregular MOPs. In light of these challenges, a refined MOEA/D algorithm utilizing infinitesimal method is proposed. This algorithm incorporates the notion of global decomposition stemming from infinitesimal method to streamline the feature information of PF, thereby facilitating the adjustment of the weight vector towards optimal distribution. Consequently, enhancements in resource allocation efficiency and algorithmic performance are achieved. In the empirical investigation, the algorithm’s performance is tested on 28 benchmarks from ZDT,DTLZ and WFG test suits.Wilcoxon’s rank-sum test and Fredman’s test were carried out on performance metrics, which proved that the proposed MOEA/D-DKS was superior to other comparison algorithms.

Development of image extraction using the centerline method in the identification of appendicitis in ultrasonography

Appendicitis is a disease that refers to inflammation of the appendix caused by obstruction, or blockage, in the lumen of the appendix. We investigated that this disease can be detected early through medical imaging such as ultrasonography (USG). However, the role of ultrasound in these cases is still limited due to the low visualization rate of the visible appendix. Based on this, this research aims to develop an image extraction process using the Centerline method in the process of identifying appendicitis in ultrasound images. The development of the extraction process is presented in the performance of the centerline and boundary extraction (CBE) algorithm which can represent image objects as boundaries that limit and separate one area from other areas. The research dataset used was 2097 ultrasound images sourced from 90 patients at the West Sumatra Lung Hospital. Based on the tests that have been carried out, it has been proven that it can reduce the width of the image object iteratively until the object is represented as a center line or the thinnest representation. The performance of the CBE algorithm in the identification process is sufficient to provide accuracy results of 92%. These results can be a new extraction concept that can provide accuracy in the identification process.

Mosquito species identification accuracy of early deployed algorithms in IDX, a vector identification tool

Mosquito-borne diseases continue to pose a great threat to global public health systems due to increased insecticide resistance and climate change. Accurate vector identification is crucial for effective control, yet it presents significant challenges. IDX - an automated computer vision-based device capable of capturing mosquito images and outputting mosquito species ID has been deployed globally resulting in algorithms currently capable of identifying 53 mosquito species. In this study, we evaluate deployed performance of the IDX mosquito species identification algorithms using data from partners in the Southeastern United States (SE US) and Papua New Guinea (PNG) in 2023 and 2024 . This preliminary assessment indicates continued improvement of the IDX mosquito species identification algorithms over the study period for individual species as well as average regional accuracy with macro average recall improving from 55.3% [Confidence Interval (CI) 48.9, 61.7] to 80.2% [CI 77.3, 84.9] for SE US, and 84.1% [CI 75.1, 93.1] to 93.6% [CI 91.6, 95.6] for PNG using a CI of 90%. This study underscores the importance of algorithm refinement and dataset expansion covering more species and regions to enhance identification systems thereby reducing the workload for human experts, addressing taxonomic expertise gaps, and improving vector control efforts.

Comparative Study on Performance of Various Neural Network Algorithms in Construction Project Cost Prediction

Comparative Study on Performance of Various Neural Network Algorithms in Construction Project Cost Prediction

Enhancing Wind Power Forecasting Accuracy with Hybrid Deep Learning and Teaching-Learning-Based Optimization

Forecasting wind power generation is crucial for ensuring grid security and the competitiveness of the power market. This paper presents an innovative approach that combines deep learning (DL) with Teaching-Learning-Based Optimization (TLBO) to predict wind power output accurately. Using a real dataset spanning diverse weather conditions and turbine specifications collected between January 2018 and March 2020, the study employs 18 features as inputs, including Ambient Temperature, Wind Direction, and Wind Speed, with real power output in kW as the target variable. Metaheuristic algorithms including Particle Swarm Optimization (PSO), Barnacles Mating Optimizer (BMO), Biogeography-Based Optimization (BBO), and Firefly Algorithm (FA) are comprehensively compared for model optimization. TLBO-DL consistently provides forecasts that closely align with actual wind power values across instances, substantiated by its low RMSE of 98.7601, indicating effective minimization of errors in wind power forecasting. Comparative analysis with other algorithms reveals that TLBO-DL outperforms PSO-DL (RMSE: 102.6627), BMO-DL (RMSE: 132.4839), BBO-DL (RMSE: 103.8517), and FA-DL (RMSE: 104.7282) in terms of overall forecasting accuracy. The variations in the performance of other algorithms across instances highlight the robustness and effectiveness of TLBO-DL in achieving accurate wind power forecasts. Overall, TLBO-DL emerges as a reliable and superior algorithm for wind power forecasting, consistently providing accurate forecasts across a range of instances.

Cloud-Cyber Physical Systems: Enhanced Metaheuristics with Hierarchical Deep Learning-based Cyberattack Detection

Cyber-Physical Systems (CPS) integrate several interconnected physical processes, networking units, and computing resources, along with monitoring the processes of the computing system. The connection between the cyber and physical world creates threatening security problems, particularly with the growing complexities of transmission networks. Despite efforts to overcome this challenge, it remains challenging to analyze and detect cyber-physical attacks in CPS. This study mainly focuses on the development of Enhanced Metaheuristics with Hierarchical Deep Learning-based Attack Detection (EMHDL-AD) method in a cloud-based CPS environment. The proposed EMHDL-AD method identifies various types of attacks to protect CPS. In the initial stage, data preprocessing is implemented to convert the input dataset into a useful format. Then, the Quantum Harris Hawks Optimization (QHHO) algorithm is used for feature selection. An Improved Salp Swarm Algorithm (ISSA) is used to optimize the hyperparameters of the HDL technique to recognize several attacks. The performance of the EMHDL-AD algorithm was examined using two benchmark intrusion datasets, and the experimental results indicated improvements over other existing approaches.

Study on progressive collapse of overall structure based on numerical simulation method and prediction of structural collapse

This study investigated the structural progressive collapse using finite element (FE) and theoretical methods. Firstly, an FE model was established based on the component method and layered shell element method. The reliability of modeling approaches was evaluated by comparing them with the previous experimental tests by the author. The NIST building was taken as an example to investigate the structural collapse behavior and the failure modes after different component failure scenarios including single column, double columns, multiple columns, and horizontal components based on the alternative path method. The results indicated that the failure of gravity columns and joints more easily triggered the progressive collapse of the structure. After the failure of double moment-resisting frame columns, the structure still demonstrated considerable robustness. The impact of floor fragments played a controlling role in the process of collapse propagation. A method for achieving rapid collapse of structures was proposed based on the structural collapse characteristics. The study found that a demolition strategy based on gravity columns can facilitate the rapid collapse of structures. After comparing the different simplified modeling methods, it was revealed that the 2D models underestimated the anti-collapse performance of structures compared to the 3D models. Under the same scenarios, the predicted structural collapse response of the 2D model was several times greater than that of the 3D model (up to 7 times). Meanwhile, the structural collapse occurred earlier in the 2D model. The anti-collapse performance of structures adopting different design codes also was compared. It was revealed that GSA 2016 has put forward requirements for higher collapse resistance of buildings. In this study, a new calculation method was proposed, namely the Collapse Probability Prediction Method Based on the Randomness of Structural Load (CPPM-RSL). In the CPPM-RSL, load reliability theory was employed to ensure that the structural collapse resistance design fell within the realm of structural engineering theory. Meanwhile, the CPPM-RSL offered greater engineering interpretability compared to previous methods. Additionally, a new simplified mechanical model was proposed to predict the collapse resistance of structures. The new simplified model did not depend on the computational results obtained from the finite element method (FEM) unlike the existing model. At the same time, the simplified method also exhibited high reliability as revealed by comparisons with the FEM. To solve highly nonlinear problems, machine learning (ML) algorithms were established to predict the anti-collapse performance and the collapse probability of structures based on the proposed theoretical model. Meanwhile, the applicability and performance of the multiple ML algorithms in estimating the structural collapse behavior were compared and evaluated. The results indicated that the LightGBM model exhibited stronger robustness and predictive ability.