Utilizing the K-Means Algorithm for Breast Cancer Diagnosis: A Promising Approach for Improved Early Detection

Abstract

Breast cancer is a pressing non-communicable disease, especially affecting women, with its incidence on the rise. In 2020, it ranked among the most common cancers in Indonesia. Timely detection and precise diagnosis are pivotal for effective breast cancer management. To enhance diagnostic accuracy, the K-means clustering method is applied to group patients based on shared attributes. This research aims to contribute significantly to breast cancer diagnosis by leveraging the K-means method, potentially improving patient survival rates.The research process involves data collection, preprocessing, K-means application, evaluation, and visualization. A dataset of 569 breast cancer patient records with 32 attributes from Kaggle is utilized. The K-Means algorithm is assessed using accuracy, yielding a value of 0.8457, signifying good performance. Malignant cases (211) and benign cases (301) are visualized in a scatter plot, distinguishing between them.In conclusion, this study presents an initial step in utilizing the K-means algorithm for breast cancer diagnosis, offering promising results. Further research and the development of more advanced models are imperative to address the global health challenge posed by breast cancer among women.Index Terms—breast cancer; clustering; K-Means Algorithm