Base Classifiers Research Articles

Stacked ensemble machine learning approach for electroencephalography based major depressive disorder classification using temporal statistics

Major depressive disorder (MDD) is a serious and widespread mental health condition that remains challenging to diagnose accurately. Traditional psychological assessments, which can be subjective and sometimes unreliable, emphasize the need for more objective diagnostic tools. In this study, we present a machine learning (ML) model designed to diagnose depression by analysing statistical time-domain features extracted from Electroencephalography (EEG) data. The model is built using a stacked ensemble ML approach, incorporating nine-base estimators with various meta-classifiers. Through multiple trials, the model achieved an accuracy of 98.01%, with precision and recall rates of 97.78% and 96.61%, respectively with Adaptive Boosting (AdaBoost) as the meta-classifer. We also investigated the effects of data sampling and the number of base classifiers on the model’s performance. The findings demonstrate that the stacked ensemble approach significantly enhances the accuracy of diagnosing MDD and that the proposed model outperforms the methods used in previous studies. This model offers a promising tool for psychologists and medical professionals to diagnose depression more reliably, potentially leading to better treatment outcomes for those affected by the disorder.

DeepBP: Ensemble deep learning strategy for bioactive peptide prediction.

Bioactive peptides are important bioactive molecules composed of short-chain amino acids that play various crucial roles in the body, such as regulating physiological processes and promoting immune responses and antibacterial effects. Due to their significance, bioactive peptides have broad application potential in drug development, food science, and biotechnology. Among them, understanding their biological mechanisms will contribute to new ideas for drug discovery and disease treatment. This study employs generative adversarial capsule networks (CapsuleGAN), gated recurrent units (GRU), and convolutional neural networks (CNN) as base classifiers to achieve ensemble learning through voting methods, which not only obtains high-precision prediction results on the angiotensin-converting enzyme (ACE) inhibitory peptides dataset and the anticancer peptides (ACP) dataset but also demonstrates effective model performance. For this method, we first utilized the protein language model-evolutionary scale modeling (ESM-2)-to extract relevant features for the ACE inhibitory peptides and ACP datasets. Following feature extraction, we trained three deep learning models-CapsuleGAN, GRU, and CNN-while continuously adjusting the model parameters throughout the training process. Finally, during the voting stage, different weights were assigned to the models based on their prediction accuracy, allowing full utilization of the model's performance. Experimental results show that on the ACE inhibitory peptide dataset, the balanced accuracy is 0.926, the Matthews correlation coefficient (MCC) is 0.831, and the area under the curve is 0.966; on the ACP dataset, the accuracy (ACC) is 0.779, and the MCC is 0.558. The experimental results on both datasets are superior to existing methods, demonstrating the effectiveness of the experimental approach. In this study, CapsuleGAN, GRU, and CNN were successfully employed as base classifiers to implement ensemble learning, which not only achieved good results in the prediction of two datasets but also surpassed existing methods. The ability to predict peptides with strong ACE inhibitory activity and ACPs more accurately and quickly is significant, and this work provides valuable insights for predicting other functional peptides. The source code and dataset for this experiment are publicly available at https://github.com/Zhou-Jianren/bioactive-peptides .

Financial Distress Prediction Using an Artificial Neural Network Integrating a Two-Stage Feature Selection Method

Financial distress prediction (FDP) is critical for companies, banks, and investors, and artificial neural networks (ANN) have been proven to be an efficient method for FDP. However, the “curse of dimensionality” in FDP not only increases the computational complexity, but also reduces the prediction accuracy. To solve this problem, this paper takes an ANN model as the basic classifier and presents a new two-stage feature selection method integrated with multiple filters and a wrapper method. The financial data of Chinese listed companies are applied for comparative analysis to verify the effectiveness of the constructed method. The results demonstrate that the proposed method achieves a smaller feature subset and better predictive effect than other methods, thus solving the “curse of dimensionality” more effectively and improving the accuracy. In addition, SHapley Additive exPlanations (SHAP) and Partial Dependence Plots (PDP) are employed to investigate the relative importance of selected features. Their results increase the credibility of the proposed model, giving users more confidence in using this “black box” model.

Multi-label Random Subspace Ensemble Classification1

In this work, we develop a new ensemble learning framework, multi-label Random Subspace Ensemble (mRaSE), for multi-label classification. Given a base classifier (e.g., multinomial logistic regression, classification tree, K-nearest neighbors), mRaSE works by first randomly sampling a collection of subspaces, then choosing the best ones that achieve the minimum cross-validation errors and, finally, aggregating the chosen weak learners. In addition to its superior prediction performance, mRaSE also provides a model-free feature ranking depending on the given base classifier. An iterative version of mRaSE is also developed to further improve the performance. A model-free extension is pursued on the iterative version, leading to the so-called Super mRaSE, which accepts a collection of base classifiers as input to the algorithm. We show the proposed algorithms compared favorably with the state-of-the-art classification algorithm including random forest and deep neural network, via extensive simulation studies and two real data applications. The new algorithms are implemented in an updated version of the R package RaSEn.

Prediction of aviation safety incident risk level based on ECSDNN

Prediction of aviation safety incident risk level is an important means of active risk management. Considering the high-dimensional complexity and class imbalance of massive aviation safety incident data, a prediction method for aviation safety incident risk level based on integrated cost-sensitive deep neural network (ECSDNN) is proposed. The feature representation of aviation safety incident data is realized by using the method of embedded coding of categorical attributes and splicing of numerical attributes; the cost-sensitive matrix and cost-sensitive loss function are designed by comprehensively considering the misclassification ratio and fixed cost, and the base classifier model based on cost-sensitive deep neural network (CSDNN) is constructed; the hard voting method is used to integrate multiple base classifiers with different parameters and performances to construct the integrated prediction model ECSDNN for aviation safety incident risk level. The experimental results on the Aviation Safety Incident Reporting System (ASRS) dataset show that compared with the optimal prediction ability of the baseline algorithm, the prediction accuracy of the ECSDNN model is improved 4.51%; compared with the single CSDNN base classifier, the prediction accuracy is improved 3.17%.

Robust Decision Support System for Stress Prediction Using Ensemble Techniques

The prevalence of stress among students has become a significant concern, impacting academic performance and overall well-being. Low-level stress can be beneficial as it helps us maintain focus on our daily routines, enabling us to perform tasks efficiently. However, severe stress can lead to heart disease, diabetes, aches, pains, and numerous other serious health issues. To address this issue, this research presents a comprehensive Decision Support System (DSS) tailored for predicting student stress levels using ensemble techniques. The goal of this project is to develop an ensemble model method of voting to classify the student dataset. An ensemble model is a combination of individual machine-learning algorithms that enhance the accuracy of the model by taking different approaches to the problem and their summary output through base classifiers. Extensive experimentation and validation are conducted using real-world student data to assess the effectiveness of the proposed DSS. In this research, we employed an ensemble modeling voting approach to classify the student dataset. Ensemble models integrate individual machine learning algorithms to enhance accuracy by leveraging the combined output of the base classifiers. Gradient Boosting Classifier (GBC), Random Forest (RF), Bagging Classifier (BC), and Extra Trees Classifier (ETC) served as base learners to construct a voting ensemble model for student dataset classification. Our findings revealed that the voting classifier consisting of Support Vector Machine (SVM), Random Forest (RF), Extra Trees Classifier (ETC), and Bagging Classifier (BC) yielded the highest accuracy at 94.78%, while GBC achieved 84.39% accuracy, RF attained 94.26% accuracy, ETC achieved 94.02% accuracy, and BC obtained 93.65% accuracy. The ensemble modeling approach significantly outperformed individual machine learning algorithms, demonstrating superior accuracy by mitigating variance and classification errors. Implementing such a system not only enhances the student assessment process in the education sector but also fosters interdisciplinary research collaboration across diverse fields within the region.

MisORFPred: A Novel Method to Mine Translatable sORFs in Plant Pri-miRNAs Using Enhanced Scalable k-mer and Dynamic Ensemble Voting Strategy.

The primary microRNAs (pri-miRNAs) have been observed to contain translatable small open reading frames (sORFs) that can encode peptides as an independent element. Relevant studies have proven that those of sORFs are of significance in regulating the expression of biological traits. The existing methods for predicting the coding potential of sORFs frequently overlook this data or categorize them as negative samples, impeding the identification of additional translatable sORFs in pri-miRNAs. In light of this, a novel method named misORFPred has been proposed. Specifically, an enhanced scalable k-mer (ESKmer) that simultaneously integrates the composition information within a sequence and distance information between sequences is designed to extract the nucleotide sequence features. After feature selection, the optimal features and several machine learning classifiers are combined to construct the ensemble model, where a newly devised dynamic ensemble voting strategy (DEVS) is proposed to dynamically adjust the weights of base classifiers and adaptively select the optimal base classifiers for each unlabeled sample. Cross-validation results suggest that ESKmer and DEVS are essential for this classification task and could boost model performance. Independent testing results indicate that misORFPred outperforms the state-of-the-art methods. Furthermore, we execute misORFPerd on the genomes of various plant species and perform a thorough analysis of the predicted outcomes. Taken together, misORFPred is a powerful tool for identifying the translatable sORFs in plant pri-miRNAs and can provide highly trusted candidates for subsequent biological experiments.

A novel ensemble approach with deep transfer learning for accurate identification of foodborne bacteria from hyperspectral microscopy

The detection of foodborne bacteria is critical in ensuring both consumer safety and food safety. If these pathogens are not properly identified, it can lead to dangerous cross-contamination. One of the most common methods for classifying bacteria is through the examination of Hyperspectral microscope imaging (HMI). A widely used technique for measuring microbial growth is microscopic cell counting. HMI is a laborious and expensive process, producing voluminous data and needing specialized equipment, which might not be widely available. Machine learning (ML) methods are now frequently utilized to automatically interpret data from hyperspectral microscopy. The objective of our study is to devise a technique that employs deep transfer learning to address the challenge of limited data and utilizes four base classifiers - InceptionResNetV2, MobileNet, ResNet101V2, and Xception - to create an ensemble-based classification model for distinguishing live and dead bacterial cells of six pathogenic strains. In order to determine the optimal weights for the base classifiers, a Powell's optimization method was utilized in conjunction with a weighted average ensemble (WAVE) technique. We carried out an extensive experimental study to verify the efficiency of our proposed ensemble model on live and dead cell images of six different foodborne bacteria. In order to gain a better understanding of the regions, we performed a Grad-CAM analysis to explain the predictions made by our model. Through a series of experiments, our proposed framework has proven its capacity to effectively and precisely detect numerous bacterial pathogens. Specifically, it achieved a perfect identification rate of 100 % for Escherichia coli (EC), Listeria innocua (LI), and Salmonella Enteritidis (SE), while achieving rates of 96.30 % for Salmonella Typhimurium (ST), 87.13 % for Staphylococcus aureus (SA), and 94.12 % for Salmonella Heidelberg (SH). As a result, it can be considered as an effective tool for the identification of foodborne pathogens, due to its high level of efficiency.

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.

BwMMV-pred: a novel ensemble learning approach using blood smear images for malaria prediction

The use of machine learning in healthcare has become widespread, enhancing the capabilities of doctors and clinicians. This study introduces a novel ensemble learning approach named Blending with Meta Majority Voting (BwMMV) for malaria prediction using blood smear images. The BwMMV technique combines the strengths of eight base classifiers to form an intermediate dataset, which is subsequently used to train five distinct meta-models using different machine learning algorithms. A Local Binary Pattern Histogram (LBPH) method is employed to extract texture features from blood smear images, effectively capturing the underlying patterns necessary for classification. The final classification decision is determined through a majority voting mechanism, selecting the outcome with the most votes as the final prediction. Our results indicate that the BwMMV approach significantly outperforms traditional hard voting and blending techniques, achieving superior accuracy, robustness, and resilience in performance. This innovative method demonstrates promising potential as a powerful tool for automated diagnosis systems, with the ability to be expanded to analyze various datasets efficiently.

Classifying zircon: A machine-learning approach using zircon geochemistry

This study presented a novel, rapid, and accurate method for determining zircon origin via a comprehensive analysis of a dataset containing 27,818 zircon trace element sets. This method integrated back propagation neural networks with the AdaBoost algorithm. The optimal classifier characterized as a linear combination of a two-layer neural network model, comprised 100 base classifiers and 400 hidden neurons. It was rigorously trained over 1000 iterations, which resulted in an unbiased error rate of 8.31%. To facilitate practical application, the classifier was integrated into a macro-enabled Excel spreadsheet.

Boosting Learning Algorithms for Chronic Diseases Prediction: A Review

Boosting algorithms are a set of machine learning techniques that are predicated on the notion that a weak learner's acquisition of multiple basic classifiers might yield results that are superior to those of any one simple classifier used alone. A comprehensive evaluation of regularly used boosting techniques against highly investigated diseases is lacking, despite the fact that boosting approaches have been used for disease prediction in many studies. Thus, the purpose of this work is to highlight the main algorithms and strategies in the boosting learning. The results of this work will help academics identify a more appropriate boosting approach to predict disease, as well as better understand current patterns and hotspots in diseases prediction models that use boosting learning. The results showed that adaboost algorithm outperformed other algorithms in terms of accuracy, achieving above 90%. This review also demonstrates how combining two boosting methods can increase the basic classifier's accuracy. By using AdaBoost and LightGBM, the accuracy reached 99.75%. XGBoost and Gradient Boosting techniques were employed more frequently in researches than other boosting algorithms.

Unified View Empirical Study for Large Pretrained Model on Cross-Domain Few-Shot Learning

The challenge of cross-domain few-shot learning (CD-FSL) stems from the substantial distribution disparities between target and source domain images, necessitating a model with robust generalization capabilities. In this work, we posit that large-scale pretrained models are pivotal in addressing the CD-FSL task owing to their exceptional representational and generalization prowess. To our knowledge, no existing research comprehensively investigates the utility of large-scale pretrained models in the CD-FSL context. Addressing this gap, our study presents an exhaustive empirical assessment of the Contrastive Language–Image Pre-Training model within the CD-FSL task. We undertake a comparison spanning six dimensions: base model, transfer module, classifier, loss, data augmentation, and training schedule. Furthermore, we establish a straightforward baseline model, E-base, based on our empirical analysis, underscoring the importance of our investigation. Experimental results substantiate the efficacy of our model, yielding a mean gain of 1.2% in 5-way 5-shot evaluations on the BSCD dataset.

Boosting certified robustness via an expectation-based similarity regularization

A certifiably robust classifier implies the one that is theoretically guaranteed to provide robust predictions against any adversarial attacks under certain conditions. Recent defense methods aim to regularize predictions by ensuring consistency across diverse perturbed samplings around the same sample, thus enhancing the certified robustness of the classifier. However, starting from the visualization of latent representations from classifiers trained with existing defense methods, we observe that noisy samplings of other classes are still easily found near a single sample, undermining the confidence in the neighborhood of inputs required by the certified robustness. Motivated by this observation, a novel training method, namely Expectation-based Similarity Regularization for Randomized Smoothing (ESR-RS), is proposed to optimize the distance between samples utilizing metric learning. To meet the requirement of certified robustness, ESR-RS focuses on the average performance of base classifier, and adopts the expected feature approximated by the average value of multiple Gaussian-corrupted samplings around every sample, to compute similarity scores between samples in the latent space. The metric learning loss is then applied to maximize the representation similarity within the same class and minimize it between different classes. Besides, an adaptive weight correlated with the classification performance is used to control the strength of the proposed similarity regularization. Extensive experiments have verified that our method contributes to stronger certified robustness over multiple defense methods without heavy computational costs.

Modified deep inductive transfer learning diagnostic systems for diabetic retinopathy severity levels classification

Diabetic Retinopathy (DR), a retinal illness that degenerates the retina and causes blindness, can be effectively treated with early detection and examination. Although expensive and unpleasant, manual retinography is the gold standard for DR diagnosis. Many Deep Learning (DL)-based algorithms have shown promise as deep learning (DR) diagnostic tools, performing similarly to human picture evaluation. These strategies work best with fine-tuned hyperparameters and huge databases. To develop Deep Inductive Transfer Learning-based Diagnostic Systems (DITL-DS) for Multiclass DR Severity Classification. The benchmark Indian Diabetic Retinopathy Image Dataset (IDRiD) is chosen first. Next, the dataset is pre-processed by proportionally scaling the image to 128*128 and using Convolutional Local Area Histogram Equalization (CLAHE). Third, the uneven class label dataset is balanced for efficiency. Fourth, a global average pooling layer and bottlenecking are added to five DITL models—Inception V3, ResNet34, EfficientNet B0, VGG16, and Xception—to compensate for information loss. Finally, we compare our improved approaches to five basic DITL classifiers. Our modified Xception model separates severity stages best, with an accuracy of 0.9840, precision of 0.9845, recall of 0.9650, AUC of 0.9979, and AUC-ROC of 0.9902. Precision, recall, and F1 scores are 0.97, 0.99, and 0.98 for stage0, 0.99, 1, and 1 for stage1, 0.98, 0.94, and 0.96 for stage2, 0.99, 0.99, and 0.99 for stage3, and 1, 1, and 1 for stage4. This may help clinicians objectively diagnose DR early. To ensure robustness, the model’s generalizability was validated using APTOS (Asia Pacific Tele-Ophthalmology Society).

Harnessing heterogeneity: A multi-embedding ensemble approach for detecting fake news in Dravidian languages

Fake news detection in low-resourced languages has become a critical challenge in our information-driven society. Contextualized word embeddings can improve the accuracy of fake news detection in low-resource languages compared to traditional word embedding methods. However, seizing the complex relationships among words in news articles is challenging when using a specific contextualized word embedding strategy. To address this issue, we propose a computationally efficient heterogeneous ensemble-based Dravidian fake news detection framework that utilizes gradient accumulation to enhance base-level and meta-level classifiers. This framework employs three optimized BiLSTM base-level classifiers, each paired with different contextualized word embeddings: multilingual BERT, XLM-RoBERTa, and MuRIL. The predictions from these base classifiers are integrated using an attention mechanism and fed into an optimized LSTM meta-classifier for final classification, ensuring a comprehensive and context-aware decision-making process. On the Dravidian_Fake dataset (Telugu, Kannada, Tamil, and Malayalam), our framework achieves significantly higher accuracy and F1 scores compared to baselines, demonstrating its effectiveness in capturing complex word relations and robustly detecting fake news in low-resource settings.

Glioma Identification Based on Digital Multimodal Spectra Integrated With Deep Learning Feature Fusion Using a Miniature Raman Spectrometer.

The miniature fiber Raman spectroscopy detection technology can reflect the properties of biomolecules through spectral characteristics and has the advantages of noninvasiveness, real-time, safety, label-free operation, and potential for early cancer diagnosis. This technology holds promise for developing portable, low-cost, intraoperative tumor detection instruments. Glioma is one of the most common malignant tumors of the central nervous system with rapid growth and a short disease course. However, the considerable heterogeneity of the glioma sample leads to substantial intraclass variance in collected spectra, coupled with the miniature Raman spectrometer's low signal-to-noise ratio. These factors diminish the accuracy of the brain glioma recognition model. To address this issue, a glioma identification method based on digital multimodal spectra integrated with deep learning features fusion (DMS-DLFF) using the miniature Raman spectrometer is proposed. Different from existing multimodal tumor detection methods employing multiple spectral instruments, DMS-DLFF enhances tumor identification accuracy without increasing hardware costs. The method mathematically decomposes the original spectra to Raman and fluorescence spectra, so as to augment the biospectral information. Then, the deep learning method is used to extract the feature information of the two kinds of spectra, respectively, and the digital multimodal spectral fusion is realized at the feature level. Moreover, a two-layer pattern recognition model is constructed based on the ensemble strategy, amalgamating the strengths of diverse classifiers. Meanwhile, the bagging strategy is introduced to improve support vector machine algorithms, one of the basic classifiers. Compared with traditional methodologies, DMS-DLFF operates at both the feature level and decision level, employing high-information-density feature vectors to train ensemble classification models for increasing overall recognition accuracy. This study collected 260 Raman spectra of glioma and 151 Raman spectra of normal brain tissue. The accuracy, sensitivity, and specificity were 91.9%, 96.7%, and 80.8%, respectively. The proposed method outperforms traditional algorithms in brain glioma detection, which helps doctors formulate precise surgical plans and thereby improve patient prognosis.

A human-in-the-loop ensemble fusion framework for road crash prediction: coping with imbalanced heterogeneous data from the driver-vehicle-environment system

ABSTRACT Road accidents are an inevitable aspect of daily life, and predicting crashes is crucial for minimizing disruptions and advancing intelligent transportation technologies. This study aims to design an ensemble fusion decision system using various base classifiers and a meta-classifier to improve crash prediction efficiency within the driver-vehicle-environment system. We adopted a data-driven strategy to analyze four categories of features—driver demographics, vehicle telemetry, driver inputs, and environmental conditions—collected from a driving simulator. Optimized modeling strategies using AdaBoost, XGBoost, GBM, LightGBM, and CatBoost were implemented. Moreover, statistical logit models were also used to assess the likelihood of crashes and the correlations among key variables. Furthermore, three resampling strategies, SMOTE-TL, SMOTE-ENN, and ADASYN, were employed to address class imbalance. The best performance was achieved with GBM, XGBoost, and AdaBoost as base classifiers, SMOTE-TL for balancing, and CatBoost as the meta-classifier, with 89.78% precision, 95.69% recall, and 92.64% F1-score.

Predicting blood lead in Uruguayan children: Individual- vs neighborhood-level ensemble learners.

Predicting childhood blood lead levels (BLLs) has had mixed success, and it is unclear if individual- or neighborhood-level variables are most predictive. An ensemble machine learning (ML) approach to identify the most relevant predictors of BLL ≥2μg/dL in urban children was implemented. A cross-sectional sample of 603 children (~7 years of age) recruited between 2009-2019 from Montevideo, Uruguay participated in the study. 77 individual- and 32 neighborhood-level variables were used to predict BLLs ≥2μg/dL. Three ensemble learners were created: one with individual-level predictors (Ensemble-I), one with neighborhood-level predictors (Ensemble-N), and one with both (Ensemble-All). Each ensemble learner comprised four base classifiers with 50% training, 25% validation, and 25% test datasets. Predictive performance of the three ensemble models was compared using area under the curve (AUC) for the receiver operating characteristic (ROC), precision, sensitivity, and specificity on the test dataset. Ensemble-I (AUC: 0.75, precision: 0.56, sensitivity: 0.79, specificity: 0.65) performed similarly to Ensemble-All (AUC: 0.75, precision: 0.63, sensitivity: 0.79, specificity: 0.69). Ensemble-N (AUC: 0.51, precision: 0.0, sensitivity: 0.0, specificity: 0.50) severely underperformed. Year of enrollment was most important in Ensemble-I and Ensemble-All, followed by household water Pb. Three neighborhood-level variables were among the top 10 important predictors in Ensemble-All (density of bus routes, dwellings with stream/other water source and distance to nearest river). The individual-level only model performed best, although precision was improved when both neighborhood and individual-level variables were included. Future predictive models of lead exposure should consider proximal predictors (i.e., household characteristics).

A new method for identification of traditional Chinese medicine constitution based on tongue features with machine learning.

The theory of Chinese medicine (TCM) constitution contributes to the optimisation of individualised healthcare programmes. However, at present, TCM constitution identification mainly relies on inefficient questionnaires with subjective bias. Efficient and accurate TCM constitution identification can play an important role in individualised medicine and healthcare. Building an efficient model for identifying traditional Chinese medicine constitutions using objective tongue features and machine learning techniques. The DS01-A device was applied to collect tongue images and extract features. We trained and evaluated five machine learning models: Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), LightGBM (LGBM), and CatBoost (CB). Among these, we selected the model with the best performance as the base classifier for constructing our heterogeneous ensemble learning model. Using various performance metrics, including classification accuracy, precision, recall, F1 score, and area under curve (AUC), to comprehensively evaluate model performance. A total of 1149 tongue images were obtained and 45 features were extracted, forming dataset 1. RF, LGBM, and CB were selected as the base learners for the RLC-Stacking. On dataset 1, RLC-Stacking1 achieved an accuracy of 0.8122, outperforming individual classifiers. After feature selection, the classification accuracy of RLC-Stacking2 improved to 0.8287, an improvement of 0.00165 compared to RLC-Stacking1. RLC-Stacking2 achieved an accuracy exceeding 0.85 for identifying each TCM constitution type, indicating excellent identification performance. The study provides a reliable method for the accurate and rapid identification of TCM constitutions and can assist clinicians in tailoring individualized medical treatments based on personal constitution types and guide daily health care. The information extracted from tongue images serves as an effective marker for objective TCM constitution identification.