Introduction

Machine learning, a branch of artificial intelligence (AI), is dedicated to developing algorithms and models that empower computers to learn from data and make informed predictions or decisions without explicit programming for each task. The fundamental concept behind machine learning is to enable systems to improve over time by learning from experience, thereby enhancing accuracy and efficiency with more data exposure. One of the most impactful applications of machine learning is in healthcare, where it plays a crucial role in disease diagnosis, treatment planning, and patient outcome prediction. In this study, we propose an optimized machine learning model specifically designed for breast cancer prediction. By leveraging data-driven learning techniques, this model aims to improve diagnostic accuracy and facilitate early detection, thereby contributing meaningfully to healthcare advancements. According to World Health Organization 2.3 million women received a breast cancer diagnosis in 2022, and 670,000 people died from the disease worldwide. In any nation on earth, breast cancer can strike women at any age after puberty, though its incidence rises with age.¹

Main Contributions of the work are as follows

Ant Colony Optimization (ACO) and machine learning techniques are combined to forecast breast cancer.

Demonstrates the efficacy of ACO in a range of machine learning applications.

Improves generalization by optimizing classifier hyperparameters and increasing prediction accuracy.

Combines a number of methods with sophisticated optimization to propose a comprehensive framework.

The proposed model incorporates a range of widely used classifiers, including K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Random Forest, Gradient Boosting, and XGBoost. These algorithms are chosen for their proven effectiveness in classification tasks and their ability to capture complex patterns within the data, making them well-suited for predictive modeling. After the initial model design, Ant Colony Optimization (ACO) is applied to fine-tune the hyperparameters, enhancing the model’s performance by optimizing key parameters. This combination of classifiers with ACO not only boosts accuracy but also strengthens the model’s reliability and robustness in identifying subtle data patterns.In swarm intelligence techniques, this algorithm is a member of the family of ant colony algorithms and is an example of a metaheuristic optimization. First put out by Marco Dorigo in his doctoral thesis in 1992.² Gradient Boosting and XGBoost are powerful ensemble techniques capable of high accuracy by successively building models to correct errors from previous iterations; Random Forest provides strong performance with resistance to overfitting; SVM is excellent at handling complex, non-linear data and high dimensional features; and KNN is straightforward and easy to understand.

The objective is to combine the capabilities of these machine learning algorithms with an optimization method that imitates ant foraging behavior to produce a more precise and effective prediction of breast cancer.Applying K Nearest Neighbor with Ant Colony Optimization produced results of 94%, Support Vector Machine with Ant Colony Optimization produced results of 96%, Gradient Boosting with Ant Colony Optimization produced results of 96%, XGBoost with Ant Colony Optimization produced results of 95.6%, and Random Forest with Ant Colony Optimization produced the highest accuracy of 99% with a false postive rate of 0.87%. Along with that AUC score , sensitivity and specificity score of the model are 98%,100% and 97% respectvely.The data collection of breast cancer from the UCI Machine Learning Repository is a valuable resource for researchers and practitioners in the field of medical diagnostics. This dataset consists of clinical and diagnostic features extracted from digitized images of breast masslesions, obtained through fine needle aspirates (FNA). With a total of 569 instances, thedataset includes 30 features, encompassing characteristics such as texture, radius, perimeter,and symmetry. Each instance is labeled as either malignant or benign, providing ground truthinformation for training and evaluating machine learning models.

The combination of various machine learning methods with the Ant Colony Optimization technique—which has received little attention in the field of breast cancer prediction, represents the work’s primary contribution. The application of  ACO has proven its effective ness in differernt machine leaning based application found in existing litearture. In a work proposed by Adamu Garba et al. application of ACO in Federated searching achieved significant improvised results.³ The use of ACO as a performance-enhancing technique, which raises prediction accuracy and optimizes classifier hyperparameters for improved generalization, is what makes this study novel. Unlike other models that rely on single classifiers or less complex optimization techniques, this methodology presents a comprehensive and resilient framework for breast cancer prediction by merging many algorithms and an advanced optimization technique³. In another work proposed by Duygu Yilmaz Eroglu et al,⁴ the application of adaptive ACO has also proven efficiency. ⁴

In section 2 few contemporary works in this field have been analyzed in details. In the section 3 and 4 the description and experimental outcomes of the proposed method have been discussed. In section 5 conclusion and the future scope of the study have been presented.

Literature Review

Numerous studies have explored machine learning approaches for cancer prediction, yielding promising results across various types of cancer. Researchers developed a breast cancer prediction model using Logistic Regression, Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Random Forest, achieving a peak accuracy of 97.14%.⁵ Another researcher applied SVM, Decision Tree, Naïve Bayes, and KNN for breast cancer prediction, reaching a maximum accuracy of 97.13% with SVM.A Scientist proposed a model utilizing Random Forest and SVM, with SVM attaining an accuracy of 86%.⁶ Scientists targeted lung cancer prediction, employing SVM, Naïve Bayes, Decision Tree, and Random Forest; Random Forest achieved the highest accuracy of 96.8%.⁷

In colorectal cancer article, researcher applied Naïve Bayes, SVM, Random Forest, and KNN, with Naïve Bayes achieving a top accuracy of 92.83%⁸. Another Scientist used Logistic Regression, Random Forest, SVM, and KNN for colorectal cancer, obtaining a peak accuracy of 98.2%⁹. For brain tumor detection, gradient vanishing and overfitting issues by implementing a transfer learning ResNet-50 model with global average pooling.¹⁰ Using Deep Neural Networks (DNN) and KNN (for k=1 and k=3), the authors reported recall, accuracy, and F-measure values of 0.97, and an AUC of 0.984 with DNN. With KNN at k=1, they achieved a recall of 0.955 and accuracy of 0.956¹¹. Their model classified images into pituitary, meninoma, and no tumor categories, yielding a general accuracy of 92.13%.¹²

Further advancements in brain tumor detection were made by who developed an MRI-based model to localize and classify tutors.¹³ In colon cancer detection article, researcher proposed a model using CNNleNetV2, which outperformed other configurations with an accuracy of 99.67%.¹⁴ In the realm of Covid-19 detection researcher utilized chest X-ray images neural networks, achieving an AUC score near 0.999, indicative of high diagnostic accuracy. Finally, Researchers classified lung and colon cancer images using the LC250 achieving 96.11% accuracy in training and 97.20% in validation for three categories: benign, adenocarcinoma, and squamous cell carcinoma. These studies collectively highlight the effectiveness of machine learning in disease prediction, ng robust performance across a variety of cancers through optimized classifiers and feature selection methods¹⁵.

Recent studies have introduced innovative machine learning¹⁶ models for disease prediction, particularly in cancer and heart disease. Scientists proposed a novel convolutional neural network, EFFI-CNN, inspired by the ICDSSPLDCNN and EASPLD-CNN trials, showcasing advancements in CNN architecture¹⁷. In heart disease prediction paper, authordeveloped a model using a concatenated ensemble classifier, achieving an accuracy of 86.89%.¹⁸ In a subsequent study, author and colleagues proposed a heart disease prediction model combining SVM-XGBoost with Particle Swarm Optimization, utilizing LASSO for feature selection, and achieving an improved accuracy of 91.8%.¹⁹

A comprehensive model presented in another study employed multiple machine learning algorithms SVM, logistic regression, decision tree classification, KNN, Gaussian Naive Bayes, and artificial neural networks reaching a high accuracy of 99%²⁰. In early-stage lung cancer detection, researcher compared several classifiers to determine the most effective technique for early diagnosis, an area particularly challenging due to subtle early symptoms. They reported an accuracy of 95.56% for SVM and 88.40% for KNN, emphasizing SVM’s effectiveness in this context. researcher’s research focused on feature selection methods, commonly utilizing Random Forest and Logistic Regression²¹. For lung cancer classification, this study tested Naive Bayes, SVM, and KNN classifiers, and achieved an accuracy of 81.25% with the Radial Basis Function (RBF) classifier²².These contributions underscore the evolving role of machine learning in healthcare²³. Enhanced feature selection, ensemble models, and optimized classifiers have significantly improved prediction accuracy in diseases with complex data patterns, offering promising tools for early diagnosis and effective intervention.

With classifiers like SVM, Random Forest, and KNN, several models achieve excellent accuracy in the literature on machine learning algorithms for cancer prediction, which shows impressive results across a variety of cancer types. Still, there are a number of obstacles and gaps. A single classifier or combinations of simple classifiers are used in many research, which may not necessarily produce the best prediction accuracy or generalizability. Furthermore, few studies integrate advanced optimization approaches like Ant Colony Optimization (ACO), even if some optimize hyperparameters or select features. Furthermore, one of the biggest challenges still facing early-stage cancer diagnosis is the issue of model resilience when dealing with complicated and unbalanced information.

By combining several potent classifiers, such as KNN, SVM, Random Forest, Gradient Boosting, and XGBoost, this suggested study fills in the gaps in the current literature and builds a more reliable and accurate predictive model for breast cancer. This method integrates many classifiers to capture a wider range of intricate patterns in the data, in contrast to earlier research that relied on single classifiers or less sophisticated optimization techniques. Furthermore, the model’s performance is improved by applying Ant Colony Optimization (ACO) for hyperparameter tweaking, which solves the optimization problem that many conventional approaches have. In comparison to current methods, this model seeks to increase diagnostic accuracy and generalization by utilizing sophisticated optimization and classifier integration, offering a more robust solution for breast cancer prediction.

Materials and Methods

This study proposes a machine learning model for breast cancer prediction, utilizing data sourced from the UCI Machine Learning Repository. Various classifiers such as K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Random Forest, Gradient Boosting, and XGBoost have been implemented to construct the predictive model. Each classifier was selected for its demonstrated effectiveness in classification tasks and ability to identify complex patterns in medical data. Following model construction, Ant Colony Optimization (ACO) was applied to fine-tune parameters, further enhancing model performance. The structural flow and processes of the proposed model are illustrated in the design diagram (Fig.1.), providing a clear overview of the integration of classifiers and optimization techniques.The data has been collected, then data preprocessing has been done, followed by applying different classifiers. After that ACO has been applied as shown in the Figure-1.This approach aims to improve diagnostic accuracy, contributing valuable insights to the field of machine learning-based disease prediction.

Figure 1: Design diagram of the proposed modelClick here to view Figure

Data Collection

The dataset utilized to train the model has been collected from the UCI data repository.²⁴ This dataset is made up of clinical and diagnostic characteristics that were taken from digital images of breast mass lesions using fine needle aspirations (FNA). Thirty parameters, including texture, radius, perimeter, and symmetry, are included in the dataset, which consists of 569 examples. Machine learning (ML) models can be trained and evaluated with ground truth information provided by the labeling of each instance as benign or malignant. To guarantee the precision and applicability of the features taken from the breast mass pictures, technicians and medical specialists worked together during the data collecting process. The primary fields in the dataset include various attributes relevant to breast cancer diagnosis. The Diagnosis field categorizes cases as either malignant (M) or benign (B). Key features include radius, representing the mean distance from the center to points on the tumor perimeter, and texture, which captures the standard deviation of gray-scale values. Additional attributes consist of perimeter and area, both fundamental to tumor size and shape assessment. The dataset also includes smoothness, indicating the local variation in radius lengths, and compactness, a measure of shape regularity. Furthermore, concavity and concave points describe the extent and number of concave regions along the tumor boundary, respectively. Symmetry and fractal dimension, the latter reflecting “coastline approximation” minus one, provide further structural information. Together, these features offer a comprehensive profile for distinguishing between malignant and benign cases.

Data Pre-processing

Thedatapre-processingmoduleisanacutecomponentinthedevelopmentofML models.

Converting Categorical filed into Numerical filed

In the above-mentioneddataset, the dependent variable diagnosis is a categorical filed representing either by ‘M’ or ‘B’. This filed has been converted into numerical value applying Label Encoding.  In machine learning, label encoding is a technique that transforms category data into numerical form so that models can process it. This approach is a straightforward and effective technique to manage categorical data since it gives each separate category a unique integer value. When categorical data contains an intrinsic order, label encoding is especially helpful since it maintains this order, which is desirable for some algorithms, such as tree-based models. Moreover, it consumes less memory and processing power than more sophisticated techniques like one-hot encoding.

Feature Selection

Another important part of building any machine learning model is feature selection. All the existing features are doing not contribute equally in decision making and sometimes presence of unnecessary feature may degrade the model’s performances. In this work SelectKBest has been applied for the feature selection.

SelectKBest, a popular machine learning feature selection strategy, helps enhance model performance by choosing the best “K” features according to their results. This approach determines which aspects are most relevant to the target variable by calculating a score of each feature and then choosing the features with highest scores. SelectKBest minimizes the dimensionality of the data by emphasizing the most significant features. This reduces noise and the chance of overfitting, resulting in more precise and broadly applicable models. SelectKBest is very helpful when working with huge datasets because it can considerably speed up the training process by minimizing the number of features. Additionally, it simplifies the generated models, which facilitates management and interpretation. SelectKBest’s adaptability to many machine learning issues stems from its capacity to operate with multiple statistical tests according on the type of data. Its popularity is further enhanced by its simplicity of implementation, which makes it possible to incorporate it effortlessly into machine learning workflows for effective and efficient feature selection.

Applying SelectKBest top 18 features have been selected for training the model. The feature importance applying SelectKBest has been shown in the Figure 2. It has been identified that perimeter_worst has the highest priority where symmetry_se has the lowest priority.

Figure 2: Feature importance applying SelectKBestClick here to view Figure

Applying Classifiers

Multiple classifiers have been applied for prediction of breast cancer as follows:

KNN Classifier

The KNN is a simple learning algorithm used for classification and regression tasks especially for instance based. It classifies data points based on the majority vote of their KNN in the feature-space.The designing of KNN model has been shown in the Table-1 below.

Table 1: Psesudo code of KNN Classifier

Support Vector Classifier (SVC)

The Support Vector Classifier (SVC) is a supervised machine learning algorithm widely applied in classification tasks. Its primary goal is to identify the optimal hyperplane that separates different classes within the feature space, ensuring maximum margin between the classes. By maximizing this margin, SVC improves classification accuracy and generalizes effectively to new data. The design process and functional architecture of SVC is illustrated in Table-2, demonstrating how the algorithm delineates the decision boundary between classes. This approach is particularly valuable in cases where a clear separation between classes is crucial for accurate predictions.

Table 2: Psesudo code of SVC Classifier

Random Forest

Random Forest is an ensemble learning technique that builds multiple decision trees during the training process and makes predictions based on the majority vote of these trees. This method employs bagging (Bootstrap Aggregating) and random feature selection, which help mitigate overfitting and enhance model accuracy. By averaging the results from diverse trees, Random Forest achieves robust performance and generalizes well to new data. Table 3 illustrates the design of the K-Nearest Neighbor (KNN) model, showcasing its approach to classification based on the proximity of data points in the feature space. Both Random Forest and KNN offer distinct, effective strategies for handling classification tasks.

Table 3: Psesudo code of SVC Classifier

XGBoost

XGBoost, or Extreme Gradient Boosting, is a highly optimized library built to enhance the gradient boosting framework. It is specifically engineered for efficient, scalable implementation of gradient boosting algorithms, renowned for delivering exceptional performance and accuracy. The design and architecture of the XGBoost model are illustrated in Table-4.

Table 4: Psesudo code of XGBoost Classifier

Gradient Boosting

Gradient Boosting is a machine learning approach that constructs an ensemble of weak learners, often decision trees, in a sequential manner. In this process, each new learner is added to the ensemble to reduce the loss function, iteratively enhancing the model’s prediction accuracy. The structure of the Gradient Boosting model is illustrated in Table-5.

Table 5: Psesudo code of Gradientboosting Classifier

Mode Optimization

For performance optimization Ant Colony Optimization has been applied. Some ant species’ foraging strategies serve as the foundation for ant colony optimization, or ACO. When these ants locate a good path, they mark it for other ants in the colony to follow by leaving pheromone markers on the ground. An analogous approach is used in ant colony optimization to address optimization issues. Within the context of swarm intelligence techniques, this algorithm belongs to the family of ant colony algorithms and represents a few metaheuristic optimizations. The first algorithm was first presented by Marco Dorigo in 1992 in his PhD thesis.^25,26 Its goal was to find the best path through a graph by mimicking the actions of ants as they navigated a path among their colony and the food source. Sample code snapshot of designing ACO has been represented in Table-6 below.

 Table 6: Psesudo code of Gradientboosting Classifier

Results

In the proposed model Ant Colony Optimization has been applied over KNN, SVM Classifier, Gradient Boosting, Random Forest, and XGBoost. It has been identified that Random Forest Classifier with Ant Colony Optimization has achieved highest accuracy of 99%. The details observations of investigational results have been shown in the Table 7 below.

Table 7: Overall experimental Outcomes of the proposed work

Confusion Matrix of the Model utilizing Random Forest and ACO has been shown in the Figure 3(a) below.

Figure 3(a): Confusion Matrix of the Model utilizing Random Forest and ACOClick here to view Figure

Figure3(b): Classification report of the Model utilizing Random Forest and ACOClick here to view Figure

The classification report has been shown in the Figure 3(b) below.

Figure 4: Oveall outcome metricsClick here to view Figure

Discussion

Combining Random Forest (RF) with Ant Colony Optimization (ACO) to achieve 99% accuracy in predictive modeling represents a major machine learning success, with several important consequences as shown in Figure 4. First, the model’s consistent performance across a range of data examples demonstrates its resilience. This indicates that the model can generalize well to new data, which increases its dependability for real-world applications. Furthermore, the model’s high prediction accuracy gives decision-makers great confidence and guarantees that the model’s outputs may be relied upon for important decisions in a variety of industries, including banking and healthcare. The model is useful for real-time and high-throughput environments because to its efficiency and scalability, which are shown by its capacity to handle massive datasets and give correct predictions in reasonable timescales. finally, the high accuracy has significant ramifications for decision-making procedures.

By providing more dependable early detection, individualized therapy, and improved predictive analytics, Random Forest with ACO models with 99% accuracy have the potential to completely transform clinical practice in the healthcare industry. Themodel has the potential to detect diseases like cancer early with high accuracy, which could result in better outcomes and earlier therapies. Additionally, it would help physicians minimize human error, optimize resources, and customize treatment approaches for each patient. Furthermore, Random Forest combined with ACO may improve decision-making and overall healthcare efficiency by more correctly predicting patient outcomes.

It can maximize results in areas such as fraud detection, resource allocation, and customized suggestions, finally resulting in better informed, significant, and impacting decisions.

The comparative analysis of the proposed work with few existing works have been shown in the Table 8.

Table 8: Comparative Analysis

Conclusion

Machine learning is a convenient tool for identification since it analyses large amounts of data. AnML based breast cancer expectation model is presented in this study. There have been several classifiers utilized, including KNN, SVM,Gradient Boosting, Random Forest, and XGBoost. To enhance performance, ACO has been applied to these models. In conclusion, the model’s superior learning and generalization abilities are demonstrated by its 99% accuracy with Ant Colony Optimization with Random Forest. It gives decision-makers an effective instrument to make highly confident and precise data-driven decisions, opening up new possibilities and promoting improvements across a range of industries. 

Acknowledgement

Authors would like to thank JIS college of Engineering, JISGROUP, for providing all kinds of R&D resources and encouraging for R&D activities.

Funding Sources

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Conflict of Interest

The authors do not have any conflict of interest.

Data Availability Statement

This statement does not apply to this article.

Ethics Statement

This research did not involve human participants, animal subjects, or any material that requires ethical approval.

Informed Consent Statement

This study did not involve human participants, and therefore, informed consent was not required.

Clinical Trial Registration

This research does not involve any clinical trials.

Permission to reproduce material from other sources

Not Applicable 

Author Contributions

Annwesha Banerjee: Conceptualization, Methodology, Writing – Original Draft.

Aniruddha Biswas and Avisekh Kumar Tiwari: Data Collection, Analysis, Writing – Review & Editing

Sumit Das: Visualization, Supervision, Project Administration. 

References

1. Breast cancer- World Health Organization-https://www.who.int/news-room/fact-sheets/detail/breast-cancer (Accessed on 23.01.2025)
2. Colorni, M. Dorigo et V. Maniezzo, Distributed Optimization by Ant Colonies, actes de la première conférenceeuropéenne sur la vie artificielle, Paris, France, Elsevier Publishing, 134-142, 1991.
3. Garba, A., Khalid, S., Aleryni, A., Ullah, I., Tairan, N.M., Shah, H. and Mumin, D. 2024. Utilizing Ant Colony Optimization for Result Merging in Federated Search. Engineering, Technology & Applied Science Research. 14, 4 (Aug. 2024), 14832–14839. DOI:https://doi.org/10.48084/etasr.7302.
   CrossRef
4. Yilmaz Eroglu, D., & Akcan, U. (2024). An Adapted Ant Colony Optimization for Feature Selection. Applied Artificial Intelligence, 38(1). https://doi.org/10.1080/08839514.2024.2335098
   CrossRef
5. Islam M, Haque M, Iqbal H, Hasan MM, Hasan M, Kabir MN. Breast Cancer Prediction: A Comparative Study Using Machine Learning Techniques. SN Comput Sci. 2020;1:290. doi:10.1007/s42979-020-00305-w
   CrossRef
6. Banerjee N, Das S. Prediction Lung Cancer– In Machine Learning Perspective. In: 2020 International Conference on Computer Science, Engineering and Applications (ICCSEA). ; 2020:1-5. doi:10.1109/ICCSEA49143.2020.9132913
   CrossRef
7. Nanda P, Duraipandian N. Prediction of Survival Rate from Non-Small Cell Lung Cancer using Improved Random Forest. In: 2020 International Conference on Inventive Computation Technologies (ICICT). ; 2020:93-97. doi:10.1109/ICICT48043.2020.9112558
   CrossRef
8. Ohata EF, Chagas JVS das, Bezerra GM, Hassan MM, de Albuquerque VHC, Filho PPR. A novel transfer learning approach for the classification of histological images of colorectal cancer. J Supercomput. 2021;77(9):9494-9519. doi:10.1007/s11227-020-03575-6
   CrossRef
9. Li H, Lin J, Xiao Y, Colorectal Cancer Detected by Machine Learning Models Using Conventional Laboratory Test Data. Technol Cancer Res Treat. 2021;20:15330338211058352. doi:10.1177/15330338211058352
   CrossRef
10. Amin J, Sharif M, Haldorai A, Yasmin M, Nayak RS. Brain tumor detection and classification using machine learning: a comprehensive survey. | Complex & Intelligent Systems | EBSCOhost. doi:10.1007/s40747-021-00563-y
    CrossRef
11. Mohsen H, El-Dahshan ESA, El-Horbaty ESM, Salem ABM. Classification using deep learning neural networks for brain tumors. Future Comput Inform J. 2018;3(1):68-71. doi:10.1016/j.fcij.2017.12.001
    CrossRef
12. Khan AH, Abbas S, Khan MA, Intelligent Model for Brain Tumor Identification Using Deep Learning. Appl Comput Intell Soft Comput. 2022;2022(1):8104054. doi:10.1155/2022/8104054
    CrossRef
13. Avşar E, Salçin K. Detection and classification of brain tumours from MRI images using faster R-CNN. Teh Glas. 2019;13(4):337-342. doi:10.31803/tg-20190712095507
    CrossRef
14. Zaki SM, Jaber MM, Kashmoola MA. Diagnosing COVID-19 Infection in Chest X-Ray Images Using Neural Network. Baghdad Sci J. 2022;19(6):1356-1356. doi:10.21123/bsj.2022.5965
    CrossRef
15. Thapa BKH Himal Chand. Lung Cancer Detection Using Convolutional Neural Network on Histopathological Images. Seventh Sense Research Group. Accessed September 30, 2024. https://dev.ijcttjournal.org//archives/ijctt-v68i10p104
16. Das S, Saha T, Nath I, Dipansu M. Exploring Machine Learning Methods for Developing a Predictive System for Parkinson’s Disease. Biosci Biotechnol Res Asia. 2024;21:569-582. doi:10.13005/bbra/3248
    CrossRef
17. Research Scholar, Acharya Nagarjuna University, Nagarjuna Nagar, Guntur, Andhra Pradesh 522510, India., Ponnada VT, Srinivasu DrSVN, Professor, Computer science and engineering, Narasaraopeta Engineering College, Narasaraopet, Andhra Pradesh 522601, India. Efficient CNN for Lung Cancer Detection. Int J Recent Technol Eng IJRTE. 2019;8(2):3499-3503. doi:10.35940/ijrte.B2921.078219
    CrossRef
18. Majumder AB, Gupta S, Singh D, Heart Disease Prediction Using Concatenated Hybrid Ensemble Classifiers. Algorithms. 2023;16(12):538. doi:10.3390/a16120538
    CrossRef
19. Majumder AB, Gupta S, Majumder S, Singh D. A Heart Disease Prediction Model using Merged XGBoost-SVM Classifier and Particle Swarm Optimization. In: 2024 5th International Conference on Mobile Computing and Sustainable Informatics (ICMCSI). ; 2024:241-248. doi:10.1109/ICMCSI61536.2024.00042
    CrossRef
20. Bhandari A, Majumder AB, Das S. An Intelligent System for Prediction of Lung Cancer Under Machine Learning Framework. In: Sharma N, Goje AC, Chakrabarti A, Bruckstein AM, eds. Data Management, Analytics and Innovation. Springer Nature; 2024:27-43. doi:10.1007/978-981-97-3242-5_3
    CrossRef
21. Abdullah D, Abdulazeez A, Sallow A. Lung cancer Prediction and Classification based on Correlation Selection method Using Machine Learning Techniques. Qubahan Acad J. 2021;1:141-149. doi:10.48161/qaj.v1n2a58
    CrossRef
22. Das S, Koley S, Saha T. Machine Learning Approaches for Investigating Breast Cancer. Biosci Biotechnol Res Asia. 2023;20:1109-1131. doi:10.13005/bbra/3163
    CrossRef
23. Das S, Nath I, Dey D, An Intelligent Approach for Detrimental Emotion Classification and Healthcare Management. In: 2024 International Conference on Computational Intelligence for Green and Sustainable Technologies (ICCIGST). ; 2024:1-5. doi:10.1109/ICCIGST60741.2024.10717562
    CrossRef
24. William Wolberg OM. Breast Cancer Wisconsin (Diagnostic). Published online 1993. doi:10.24432/C5DW2B
25. Colorni A, Dorigo M, Maniezzo V. Distributed Optimization by Ant Colonies.; 1991.
26. Dorigo M. Optimization, learning and natural algorithms. Ph Thesis Politec Milano. Published online 1992. Accessed September 30, 2024. https://cir.nii.ac.jp/crid/1573950400977139328
27. Apoorva V, Yogish HK, Chayadevi ML. Breast Cancer Prediction Using Machine Learning Techniques. In: Atlantis Press; 2021:348-355. doi:10.2991/ahis.k.210913.043
    CrossRef