1Research Scholar, Biju Patnaik University of Technology, Rourkela, Odisha, India
Email: for_manumohan@yahoo.co.in
2Principal, Synergy Institute of Technology, Phulnakhara, Bhubaneswar, India
Email: asw_moh@yahoo.com
Globally, breast cancer ranks as the second leading cause of cancer-related mortality among women, with a significant impact on middle-aged populations. Early detection and prevention play a vital role in reducing mortality, and accurate prognosis—along with the ability to predict recurrence risk—is essential for effective disease management.
In this study, we focus on improving breast cancer prediction and detection accuracy by applying a set of classification algorithms after optimizing the dataset through a combined feature selection strategy. This research is based on the Wisconsin Breast Cancer Dataset (WBCD), which was accessed via the UCI Machine Learning Repository and used as the principal dataset. A total of five classification models—Random Forest, Decision Tree, Logistic Regression, k-NN, and AdaBoost—were implemented to analyze 35 variables and measure their predictive capability.
To enhance these models, two complementary feature selection algorithms were applied separately to identify the most informative features. The top-ranked features from both methods were then merged into a unified set, while low-importance attributes were removed. Eliminating less relevant features not only streamlined the data but also reduced noise, ultimately leading to notable improvements in accuracy across all five classification approaches.
Keywords: Breast cancer detection, Data mining algorithms, Feature selection, Wisconsin Breast Cancer Dataset, Random Forest, Decision Tree, Logistic Regression, k-NN, AdaBoost, Optimization algorithms.
How to cite this article: Sahoo M, Mohanty AK. Performance Analysis of Different Data Mining Algorithms Using Combined Features from Two Optimization Algorithms for Breast Cancer Detection. Int J Drug Deliv Technol. 2026;16(7s): 764-770; DOI: 10.25258/ijddt.16.7s.82
Source of support: None
Conflict of interest: None