Ensemble Machine Learning Algorithms for Precision Breast Cancer Diagnosis: A Multi-criteria Evaluation Approach

• Published in: SN computer science Vol. 6; no. 2; p. 153
• Main Authors: Pallapu, Srinivasa Rao, Syed, Khasim
• Format: Journal Article
• Language: English
• Published: Singapore Springer Nature Singapore 11.02.2025; Springer Nature B.V
• Subjects: Algorithms; Artificial neural networks; Bioinformatics; Collaboration; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data augmentation; Data Structures and Information Theory; Datasets; Decision trees; Diagnosis; Ensemble learning; Illnesses; Information Systems and Communication Service; Machine learning; Medical imaging; Medical research; Multiple criterion; Neural networks; Original Research; Pattern Recognition and Graphics; Research Advancements in Intelligent Computing; Research projects; Software Engineering/Programming and Operating Systems; Support vector machines; Teams; Vision; United States > US; Wisconsin; CNN; Breast cancer; SMOTE; DICOM; Ensemble learning
• ISSN: 2662-995X; 2661-8907
• DOI: 10.1007/s42979-025-03676-0

Subsequent researchers and medical teams have reported that cancer is a diverse illness with several subtypes. Such cancer-causing illnesses are to be diagnosed at an early stage and the prognosis leads to various research among the cancer community. Because of the importance of placing those with cancer as high- or low-risk, several research teams in the biomedical and bioinformatics areas have investigated the use of machine-learning approaches. The ultimate goal of the research project is to create an algorithm for multi-criteria ensemble learning. to detect breast cancer. Reducing the amount of time required to process the classifier is the ultimate goal of the study, which will increase accuracy. The system takes the DICOM images and pre-processes them on applying SMOTE and data augmentation. The data is then analysed for feature extraction through a convolution neural network. The optimised DICOM images are fed as input to ensemble learning algorithms framed over Ada-boost(CNNbAB), gradient-boost decision tree (CNNbGBDT) and random forest (CCNbRF) algorithms. The reason behind the choice of the algorithm is to simplify the processing and time complexity of the time without compromising the accuracy of CNN. It is observed that the accuracy of the system is 98.65% which is found to be a better value than the existing system for cancer diagnosis.