An Improved Diagnostic Classification of Breast Cancer Using Ensemble and Deep Learning Methods

• Ashish Gupta,
• Deepak Gupta,
• Deepali Virmani,
• Rolly Gupta,
• Ashwin. M,
• John Philip Bhimavarapu,

Vol. 19, No. 12, December 31, 2025
10.3837/tiis.2025.12.001, Download Paper (Free):

• breast cancer
• Voting ensemble classifier
• Classification
• Deep Learning

Abstract

Over the previous ten years, breast cancer has continuously been the highly prevalent kind of cancer among women. Several approaches have been proposed for the detection of breast cancer (BC). Concentrating mostly on the differentiation and classification of benign and malignant tumors. Machine learning (ML) has appeared as a tool of experimentation for breast cancer disease classification, with data mining as well as classification techniques proving efficient in prediction and categorization. Among these methods, ensemble-based classification has shown optimal results, leveraging multiple approaches to achieve the most accurate outcomes. In this research, we utilised the UC Irvine Breast Cancer (UCI) dataset and BRCA to develop a voting ensemble classifier that integrates 4 distinct ML techniques: Extra Trees Classifier (ETC), Light Gradient Boosting Machine (LGBM), Naive Bayes classifier (NB), as well as Linear Discriminant Analysis (LDA). The suggested ELNL-T achieved outstanding performance metrics, including an accuracy rate of 97.71%, Precision of 96.82%, sensitivity of 99.57%, and F1 score of 98.61%. Various metrics were used to evaluate our model efficacy and efficiency, alongside comparisons with individual classifiers and established cutting edge techniques. The main goal of this research is to find the highly significant ensemble ML classifier for identification as well as diagnosis of breast cancer and focusing on accuracy along with AUC score.

Statistics

Cite this article