Voting Classifier Research Articles

Abstract 4113047: Discovering novel biomarkers and predicting cardiovascular disease using AI/ML techniques for precision medicine

Personalized interventions are deemed vital given the intricate characteristics, advancement, inherent genetic composition, and diversity of cardiovascular diseases (CVDs). The appropriate utilization of artificial intelligence (AI) and machine learning (ML) methodologies can yield novel understandings of CVDs, enabling improved personalized treatments through predictive analysis and deep phenotyping. In this study, we proposed and employed a novel approach combining traditional statistics and a nexus of cutting-edge AI/ML techniques to identify significant biomarkers for our predictive engine by analyzing the complete transcriptome of CVD patients (Figure 1). After robust gene expression data pre-processing, we utilized three statistical tests to assess the differences in transcriptomic expression and clinical characteristics between healthy individuals and CVD patients. Next, a classifier assigned rankings to transcriptomic features based on their relation to the case-control variable. The top ten percent of commonly observed significant biomarkers were evaluated using four unique ML classifiers. After optimizing hyperparameters, the ensembled models, which were implemented using a soft voting classifier, accurately differentiated between patients and healthy individuals. We have uncovered 18 transcriptomic biomarkers that are highly significant in the CVD population that were used to predict disease with up to 96% accuracy (Figure 2). Additionally, we cross-validated our results with clinical records collected from patients in our cohort (Figure 3). The identified biomarkers served as potential indicators for early detection of CVDs. With its successful implementation, our newly developed predictive engine provides a valuable framework for identifying patients with CVDs based on their biomarker profiles.

Advancing ischemic stroke diagnosis and clinical outcome prediction using improved ensemble techniques in DSC-PWI radiomics.

Ischemic stroke is a leading global cause of death and disability and is expected to rise in the future. The present diagnostic techniques, like CT and MRI, have some limitations in distinguishing acute from chronic ischemia and in early ischemia detection. This study investigates the function of ensemble models based on the dynamic radiomics features (DRF) from the dynamic susceptibility contrast perfusion-weighted imaging (DSC-PWI) ischemic stroke diagnosis, neurological impairment assessment, and modified Rankin Scale (mRS) outcome prediction). DRF is extracted from the 3D images, features are selected, and dimensionality is reduced. After that, ensemble models are applied. Two model structures were developed: a voting classifier with 6 bagging classifiers and a stacking classifier based on 4 bagging classifiers. The ensemble models were evaluated on three core tasks. The Stacking_ens_LR model performed best for ischemic stroke detection, the LR Bagging model for NIH Stroke Scale (NIHSS) prediction, and the NB Bagging model for outcome prediction. These outcomes illustrate the strength of ensemble models. The work showcases the role of ensemble models and DRF in the stroke management process.

Optimizing SMS Spam Detection Using Machine Learning: A Comparative Analysis of Ensemble and Traditional Classifiers

With the rapid rise of mobile communication, Short Message Service (SMS) has become an essential platform for transmitting information. However, the growing volume of unsolicited and harmful spam messages presents significant challenges for both users and mobile network operators. This study explores the effectiveness of various machine learning models, including Random Forest, Gradient Boosting, AdaBoost, Support Vector Machine (SVM), Logistic Regression, and an Ensemble Voting Classifier, in detecting SMS spam. A dataset containing 5,572 SMS messages, labeled as either spam or ham (legitimate), was used to evaluate these models. Hyperparameter tuning was performed on each model to optimize accuracy, and the models were assessed using metrics such as precision, recall, F1-score, and accuracy. The results indicated that the SVM and Ensemble Voting Classifier achieved the highest performance, with accuracies of 0.9857 and 0.9848, respectively. Both models demonstrated superior recall for spam messages, making them highly effective for real-world spam detection systems. While Random Forest, Gradient Boosting, and AdaBoost also performed well, their slightly lower recall for spam suggests that they may misclassify some spam as legitimate messages. The study highlights the effectiveness of machine learning models in addressing the SMS spam problem, particularly when using ensemble methods. Future research should focus on addressing class imbalance and exploring deep learning approaches to further enhance model performance. These findings offer valuable insights for developing more accurate and scalable SMS spam detection systems.

Enhancing the accuracy of blockage detection in centrifugal pump using the majority voting classifier on an unbalanced dataset

Blockage in a centrifugal pump can adversely impact its performance. In this study, an experimental facility is developed to simulate three types of blockages (suction, discharge, and simultaneous suction and discharge). To classify these faults and detect their severity, a methodology involving the application of majority voting classifier (MVC) to the pump’s discharge pressure signals is presented. An unbalanced dataset is constructed, where the number of samples for a specific blockage condition decreases with increasing fault severity. Statistical features, entropy features, and entropies meta features are extracted from the signal and ranked using XGBoost and minimum redundancy maximum relevance (MRMR). Subsequently, the optimal features are selected based on the best performing model (Linear Discriminant Analysis) among ten different models. Best four models are selected and ensembled to form MVC. Results show that MVC achieves an accuracy of 89.90% and 88.26% for features selected by XGBoost and MRMR, respectively. Finally, the unbalanced dataset is balanced using synthetic minority oversampling and it is shown that MVC achieves an accuracy of 100% on this balanced dataset for the features selected using both approaches.

Efficient Deep Learning Methods for Detecting Road Accidents by Analyzing Traffic Accident Images

Speed is one of the major factors in car crashes. Many lives could have been saved if emergency services had been alerted to the disaster and arrived in time. For the sake of protecting valuable human lives, an effective automatic accident detection system with prompt reporting of the accident scene to emergency services is essential. Therefore, this research proposes some effective Deep Learning techniques that properly recognize the incidence of accidents. The paper introduces two different techniques for image classification, with a particular focus on distinguishing between accident and non-accident images. The dataset used for the proposed model is taken from Kaggle, which is a collection of CCTV images. The first approach is a hybridized TL-ML method that employs transfer learning techniques that use different pre-trained versions of convolutional neural networks to extract features from image datasets. These extracted features are then fed into various machine learning classifiers to categorize the images as either Accident or Non-accident. To make the final decision, a voting classifier is utilized to choose the best classification outcome from the set of previously employed machine learning classifiers. In the second method, a modified Convolutional Neural Network (CNN) called SpinalNet is adopted. The performance of these models was evaluated by comparing them with each other and with a customized CNN base model. SpinalNet consistently surpassed the other models in terms of Precision, Recall, F1-Score, and Accuracy, demonstrating its outstanding capabilities.

Classification of an Individual's Vaccination Status Using Ensemble Hard Voting Classifier

Vaccination is a proactive medical immunization procedure where an inactivated form of a disease-causing agent (such as a virus) is administered to boost the body's defense systems. Efficient management of vaccination status is crucial in healthcare management, disease eradication, community immunity ("herd immunity"), disease prevention, and global health security. Ensuring precise monitoring and validation of an individual's vaccination status is indispensable, especially in the context of emerging diseases and epidemics. This study evaluates the likelihood of individuals obtaining vaccination for the H1N1 virus and the seasonal flu vaccine. Ensemble methods combine the predictions of multiple base classifiers to enhance overall performance. One such method, the hard voting classifier, aggregates the votes from each base classifier and selects the class with the majority vote as the final prediction. This approach leverages the strengths of different classifiers, reducing the risk of individual model biases and improving generalization using metrics such as precision, recall, accuracy, and F1-score are employed to assess the system's effectiveness. The results demonstrate how data-driven methods can address population wellness and improve vaccination rates using an ensemble method. The proposed ensemble hard voting classifier achieved accuracies of 0.905 and 0.907 on the H1N1 and seasonal vaccine datasets, respectively. Using an ensemble approach like the hard voting classifier enhances prediction accuracy and robustness, ultimately leading to better decision making in public health initiatives.

Hybridization of Machine Learning Models for Alzheimers Disease Classification

Alzheimer's disease (AD), is a gradual cognitive decline and memory impairment. It is a major health concern worldwide. Despite intensive research efforts, accurate and early diagnosis remains difficult to achieve, largely due to the complexity of AD pathology and the absence of definitive biomarkers. Existing diagnostic approaches often rely on costly and invasive procedures, leading to delays in diagnosis and treatment initiation, and limiting the effectiveness of therapeutic interventions. To overcome these issues, this work suggests a novel approach for AD classification using EEG signals. EEG signals offer a non-invasive and cost-effective means of assessing brain activity, making them an attractive candidate for biomarker discovery and disease classification. The proposed work integrates preprocessing, feature extraction, and classification methodologies to accurately differentiate between AD, normal/healthy states, and Frontotemporal Dementia (FTD). The proposed solution begins with Sequential Savitzky-Golay filtering (SEQ-SG) to enhance the quality of EEG signals by reducing noise and enhancing relevant features. Subsequently, an Improved Principal Component Analysis (IPCA) approach is employed for feature extraction, incorporating feature scaling using StandardScaler to ensure uniform contribution from all features. Finally, classification is achieved using a hybrid approach named HMLCAD (Hybridization of Machine Learning for Classification of Alzheimer's Disease), which combines Random Forest and Gradient Boosting through a voting classifier ensemble. This methodology offers a promising framework for accurate and early detection of AD, enabling timely intervention and improved patient outcomes.

Optimizing breast cancer diagnosis: Harnessing the power of nature-inspired metaheuristics for feature selection with soft voting classifiers

Breast cancer is a widespread and serious condition that poses a significant threat to women's health globally, contributing significantly to mortality rates. Machine learning tools play a critical role in both the effective management and early detection of this disease. Feature selection (FS) methods are essential for identifying the most impactful features to improve breast cancer diagnosis. These methods reduce data dimensionality, eliminate irrelevant information, enhance learning accuracy, and improve the comprehensibility of results. However, the increasing complexity and dimensionality of cancer data pose substantial challenges to many existing FS methods, thereby reducing their efficiency and effectiveness. To overcome these challenges, numerous studies have demonstrated the success of nature-inspired optimization (NIO) algorithms across various domains. These algorithms excel in mimicking natural processes and efficiently solving complex optimization problems. Building on these advancements, we propose an innovative approach that combines powerful feature selection methods based on NIO techniques with a soft voting classifier. The NIO techniques employed include the Genetic Algorithm, Cuckoo Search, Salp Swarm, Jaya, Flower Pollination, Whale Optimization, Sine Cosine, Harris Hawks, and Grey Wolf Optimization algorithms. The Soft Voting Classifier integrates various machine learning models, including Support Vector Machines, Gaussian Naive Bayes, Logistic Regression, Decision Tree, and Gradient Boosting. These are used to improve the effectiveness and accuracy of breast cancer diagnosis. The proposed approach has been empirically evaluated using a variety of evaluation measures, such as F1 score, precision, recall, accuracy and Area Under the Curve (AUC), for performance comparison with individual machine learning techniques. The results demonstrate that the soft-voting ensemble technique, particularly when combined with feature selection based on the Jaya algorithm, outperforms all individual classifiers on the breast cancer dataset. It achieves the highest scores for accuracy, precision, recall, F1-score, and AUC, with values of 99.6 %, 99.21 %, 100 %, 99.6 %, and 99.6 %, respectively. Following closely, both the Genetic Algorithm and Salp Swarm feature selection demonstrate strong performance, with scores of 99.5 %, 100 %, 99 %, 99.5 %, and 99.5 %, respectively. Furthermore, the approach's effectiveness in identifying the minimal set of relevant features has been evaluated. Given the critical importance of selecting relevant features in complex cancer datasets, the proposed method proves to be a valuable tool for achieving early detection and improved interpretability. This method can also support healthcare professionals in making well-informed decisions regarding cancer diagnosis and treatment strategies.

Diabetic retinopathy data augmentation and vessel segmentation through deep learning based three fully convolution neural networks

ProblemThe eye fundus imaging is used for early diagnosis of most damaging concerns such as diabetic retinopathy, retinal detachments and vascular occlusions. However, the presence of noise, low contrast between background and vasculature during imaging, and vessel morphology lead to uncertain vessel segmentation. AimThis paper proposes a novel retinalblood vessel segmentation method for fundus imaging using a Difference of Gaussian (DoG) filter and an ensemble of three fully convolutional neural network (FCNN) models. MethodsA Gaussian filter with standard deviation σ1 is applied on the preprocessed grayscale fundus image and is subtracted from a similarly applied Gaussian filter with standard deviation σ2 on the same image. The resultant image is then fed into each of the three fully convolutional neural networks as the input. The FCNN models' output is then passed through a voting classifier, and a final segmented vessel structure is obtained.The Difference of Gaussian filter played an essential part in removing the high frequency details (noise) and thus finely extracted the blood vessels from the retinal fundus with underlying artifacts. ResultsThe total dataset consists of 3832 augmented images transformed from 479 fundus images. The result shows that the proposed method has performed extremely well by achieving an accuracy of 96.50%, 97.69%, and 95.78% on DRIVE, CHASE,and real-time clinical datasets respectively. ConclusionThe FCNN ensemble model has demonstrated efficacy in precisely detecting retinal vessels and in the presence of various pathologies and vasculatures.

Supervised learning via ensembles of diverse functional representations: the functional voting classifier

Supervised learning via ensembles of diverse functional representations: the functional voting classifier

Late feature fusion using neural network with voting classifier for Parkinson’s disease detection

Parkinson’s disease (PD) is classified as a neurological, progressive illness brought on by cell death in the posterior midbrain. Early PD detection will assist doctors in reducing the disease’s consequences. A collection of skilled models that may be applied to regression as well as classification is known as artificial intelligence (AI). PD can be detected using a variety of dataset formats, including text, speech, and picture datasets. For the purpose of classifying Parkinson’s disease, this study suggests merging deep with machine learning recognition approaches. The three primary components of the suggested approach are designed to enhance the accuracy of Parkinson’s disease early diagnosis. These sections cover the topics of categorising, combining, and separating. Convolutional Neural Networks (CNN) as well as attention procedures are used to create feature extractors. The related motion signals are fed to a combination of convolutional neural network and long-short-memory model for feature extraction. Besides, for the classification of patients from non-suffers of Parkinson’s disease, Random Forest, Logistic Regression, Support Vector Machine, Extreme Boot Classifier, and voting classifier were used. Our result shows that for the PD handwriting and related motion datasets, using the proposed CNN with an attention and voting classifier yields 99.95% accuracy, 99.99% precision, 99.98% sensitivity, and 99.95% F1-score. Based on these results, it is warranted to conclude that the proposed methodology of feature extraction from photos of handwriting and relating motor symptoms, fusing of those features, and following it with a voting classifier yields excellent results for PD classification.

Application of machine learning classification algorithms to predict probability of channel presence from seismic data

The article analyzes the application of various classification algorithms (Random Forest, Extra Trees, Gradient Boosting, Multi-layer Perceptron, Gaussian Naive Bayes, Voting Classifier) based on supervised machine learning, both for individual seismic attributes (amplitude, phase, frequency) and for a set of seismic attributes calculated in the interval of channel system development for probabilistic assessment of presence of channels. The object of the study is the deposits of the Tyumen suite of a group of deposits located in the Khanty-Mansiysk Autonomous Okrug. The purpose of this work is to find the most effective seismic attributes and machine learning classification methods to optimize the methodology and improve the accuracy of forecasting the probability of channel system presence.

An automated approach to predict diabetic patients using KNN imputation and effective data mining techniques

Diabetes is thought to be the most common illness in underdeveloped nations. Early detection and competent medical care are crucial steps in reducing the effects of diabetes. Examining the signs associated with diabetes is one of the most effective ways to identify the condition. The problem of missing data is not very well investigated in existing works. In addition, existing studies on diabetes detection lack accuracy and robustness. The available datasets frequently contain missing information for the automated detection of diabetes, which might negatively impact machine learning model performance. This work suggests an automated diabetes prediction method that achieves high accuracy and effectively manages missing variables in order to address this problem. The proposed strategy employs a stacked ensemble voting classifier model with three machine learning models. and a KNN Imputer to handle missing values. Using the KNN imputer, the suggested model performs exceptionally well, with accuracy, precision, recall, F1 score, and MCC of 98.59%, 99.26%, 99.75%, 99.45%, and 99.24%, respectively. In two scenarios one with missing values eliminated and the other with KNN imputer, the study thoroughly compared the suggested model with seven other machine learning techniques. The outcomes demonstrate the superiority of the suggested model over current state-of-the-art methods and confirm its efficacy. This work demonstrates the capability of KNN imputer and looks at the problem of missing values for diabetes detection. Medical professionals can utilize the results to improve care for diabetes patients and discover problems early.

Predicting compressed earth blocks compressive strength by means of machine learning models

Compressed Earth Blocks (CEB) are an interesting alternative to conventional masonry units. They are unfired and provide a thermal comfort in the constructions. However, their compressive strength needs to be assessed to ensure a mechanical stability. The latter depends on the soil variability, as well as several manufacturing parameters such as water content, compaction pressure, stabilizer type and proportion. This is challenging as it requires time and effort to adapt the parameters to achieve satisfactory results. In this study, machine learning classification models were trained using historical data for predicting CEB compressive strength. Voting Classifier (VC) provided the highest performance with an accuracy of 78 %. SHapley Additive exPlanations (SHAP) were used to identify and prioritize the features in the model's decision-making process. The compaction pressure and soil granularity were the most decisive parameters. VC was also tested to assess the compressive strength of sediment-based CEB manufactured at the laboratory scale.

Deep learning with ensemble approach for early pile fire detection using aerial images

Wildfires rank among the world’s most devastating and expensive natural disasters, destroying vast forest resources and endangering lives. Traditional firefighting methods, reliant on ground crew inspections, have notable limitations and pose significant risks to firefighters. Consequently, drone-based aerial imaging technologies have emerged as a highly sought-after solution for combating wildfires. Recently, there has been growing research interest in autonomous wildfire detection using drone-captured images and deep-learning algorithms. This paper introduces a novel deep-learning-based method, distinct in its integration of infrared thermal, white, and night vision imaging to enhance early pile fire detection, thereby addressing the limitations of existing methods. The study evaluates the performance of machine learning algorithms such as random forest (RF) and support vector machines (SVM), alongside pre-trained deep learning models including AlexNet, Inception ResNetV2, InceptionV3, VGG16, and ResNet50V2 on thermal-hot, green-hot, and white-green-hot color images. The proposed approach, particularly the ensemble of ResNet50V2 and InceptionV3 models, achieved over 97% accuracy and over 99% precision in early pile fire detection on the FLAME dataset. Among the tested models, ResNet50V2 excelled with the thermal-fusion palette, InceptionV3 with the white-hot and green-hot fusion palettes, and VGG16 with a voting classifier on the normal spectrum palette dataset. Future work aims to enhance the detection and localization of pile fires to aid firefighters in rescue operations.

A Robust Solution for Tea Leaf Classification using the SMOTE and Soft Voting Classifier

Background/Objectives: The purpose of the proposed work is to identify how a tea leaf is conceptualized. This further enables the categorization of leaves into healthy and unhealthy groups, assisting both the vendor and customer in obtaining the right quality for their requirements. The primary cause of the low production of tea in various parts of the country is the late detection of diseased leaves. Literature suggests that the current AI techniques can detect but have a high level of similarities. Hence, the objective is to identify diseased leaf images at early stage. Methods: Earlier studies have used mostly CNN for classification which is more time consuming, and they create their own parameters. In this study, we extracted the textual features of images by using GLCM, and further classification was implemented with the help of SMOTE and a soft voting classifier. Findings: The proposed method resulted in optimal timeline without using CNN for classification. The accuracy is better when compared to current methods. Novelty: This work presents a novel approach in calculating textual features of leaves that can be further used for classification or tea-leaf grading purposes. Keywords: Image processing, Feature extraction, GLCM, SMOTE, Voting classifier

Predictive modelling of metabolic syndrome in Ghanaian diabetic patients: an ensemble machine learning approach

Abstract Objectives The burgeoning prevalence of cardiometabolic disorders, including type 2 diabetes mellitus (T2DM) and metabolic syndrome (MetS) within Africa is concerning. Machine learning (ML) techniques offer a unique opportunity to leverage data-driven insights and construct predictive models for MetS risk, thereby enhancing the implementation of personalised prevention strategies. In this work, we employed ML techniques to develop predictive models for pre-MetS and MetS among diabetic patients. Methods This multi-centre cross-sectional study comprised of 919 T2DM patients. Age, gender, novel anthropometric indices along with biochemical measures were analysed using BORUTA feature selection and an ensemble majority voting classification model, which included logistic regression, k-nearest neighbour, Gaussian Naive Bayes, Gradient boosting classification, and support vector machine. Results Distinct metabolic profiles and phenotype clusters were associated with MetS progression. The BORUTA algorithm identified 10 and 16 significant features for pre-MetS and MetS prediction, respectively. For pre-MetS, the top-ranked features were lipid accumulation product (LAP), triglyceride-glucose index adjusted for waist-to-height ratio (TyG-WHtR), coronary risk (CR), visceral adiposity index (VAI) and abdominal volume index (AVI). For MetS prediction, the most influential features were VAI, LAP, waist triglyceride index (WTI), Very low-density cholesterol (VLDLC) and TyG-WHtR. Majority voting ensemble classifier demonstrated superior performance in predicting pre-MetS (AUC = 0.79) and MetS (AUC = 0.87). Conclusion Identifying these risk factors reveals the complex interplay between visceral adiposity and metabolic dysregulation in African populations, enabling early detection and treatment. Ethical integration of ML algorithms in clinical decision-making can streamline identification of high-risk individuals, optimize resource allocation, and enable precise, tailored interventions.

Improving Cleaning of Solar Systems through Machine Learning Algorithms

The study focuses on the importance of maintaining photovoltaic (PV) systems for optimal performance in sustainable energy generation. It highlights the impact of dust accumulation on reducing system efficiency and proposes a method to predict system performance, aiding in scheduling cleaning activities effectively. Two prediction models are developed: one using time-series prediction techniques (LSTM, ARIMA, SARIMAX) to forecast Performance Ratio (PR), and another employing ensemble voting classifiers (RF, Log, GBM) to predict the need for cleaning. The SARIMAX model performs best, achieving high accuracy in PR prediction (R2 = 92.12%), while the classification model accurately predicts cleaning needs (91%). The research provides valuable insights for improving maintenance strategies and enhancing the efficiency and sustainability of PV systems.

Mitigating Sociodemographic Bias in Opioid Use Disorder Prediction: Fairness-Aware Machine Learning Framework.

Opioid use disorder (OUD) is a critical public health crisis in the United States, affecting >5.5 million Americans in 2021. Machine learning has been used to predict patient risk of incident OUD. However, little is known about the fairness and bias of these predictive models. The aims of this study are two-fold: (1) to develop a machine learning bias mitigation algorithm for sociodemographic features and (2) to develop a fairness-aware weighted majority voting (WMV) classifier for OUD prediction. We used the 2020 National Survey on Drug and Health data to develop a neural network (NN) model using stochastic gradient descent (SGD; NN-SGD) and an NN model using Adam (NN-Adam) optimizers and evaluated sociodemographic bias by comparing the area under the curve values. A bias mitigation algorithm, based on equality of odds, was implemented to minimize disparities in specificity and recall. Finally, a WMV classifier was developed for fairness-aware prediction of OUD. To further analyze bias detection and mitigation, we did a 1-N matching of OUD to non-OUD cases, controlling for socioeconomic variables, and evaluated the performance of the proposed bias mitigation algorithm and WMV classifier. Our bias mitigation algorithm substantially reduced bias with NN-SGD, by 21.66% for sex, 1.48% for race, and 21.04% for income, and with NN-Adam by 16.96% for sex, 8.87% for marital status, 8.45% for working condition, and 41.62% for race. The fairness-aware WMV classifier achieved a recall of 85.37% and 92.68% and an accuracy of 58.85% and 90.21% using NN-SGD and NN-Adam, respectively. The results after matching also indicated remarkable bias reduction with NN-SGD and NN-Adam, respectively, as follows: sex (0.14% vs 0.97%), marital status (12.95% vs 10.33%), working condition (14.79% vs 15.33%), race (60.13% vs 41.71%), and income (0.35% vs 2.21%). Moreover, the fairness-aware WMV classifier achieved high performance with a recall of 100% and 85.37% and an accuracy of 73.20% and 89.38% using NN-SGD and NN-Adam, respectively. The application of the proposed bias mitigation algorithm shows promise in reducing sociodemographic bias, with the WMV classifier confirming bias reduction and high performance in OUD prediction.

A two-stage ensemble learning based prediction and grading model for PD-1/PD-L1 inhibitor-related cardiac adverse events: A multicenter retrospective study

BackgroundImmune-related cardiac adverse events (ircAEs) caused by programmed cell death protein-1 (PD-1) and programmed death-ligand-1 (PD-L1) inhibitors can lead to fulminant and even fatal consequences. This study aims to develop a prediction and grading model for ircAEs, enabling graded management of patients. MethodsThis study utilized medical record systems from two medical institutions to develop a prediction and grading model for ircAEs using ten machine learning algorithms and two variable screening methods. The model was developed based on a two-stage ensemble learning framework. In the first stage, the ircAEs and non-ircAEs cases were classified. In the second stage, ircAEs cases were grouped into grades 1–2 and 3–5. The experiments were evaluated using five-fold cross-validation. The model's prediction performance was assessed using accuracy, precision, recall, F1 value, Brier score, receiver operating characteristic curve area (AUC), and area under the precision-recall curve (AUPR). Results615 patients were included in the study. 147 experienced ircAEs, and 44 experienced grade 3–5 ircAEs. The soft voting classifier trained using the variables screened by feature importance ranking performed better than other classifiers in both stages. The average AUC for the first and second stages is 84.18 % and 85.13 %, respectively. In the first stage, the three most important variables are N-terminal B-type natriuretic peptide (NT-proBNP), interleukin-2 (IL-2), and C-reactive protein (CRP). In the second stage, the patient's age, NT-proBNP, and left ventricular ejection fraction (LVEF) are the three most critical variables. ConclusionsThe prediction and grading model of ircAEs based on two-stage ensemble learning established in this study has good performance and potential clinical application.