Work

Supervised Machine Learning-Based Classification and Prediction of Breast Cancer

Year: 2024
Type: article
Abstract: There is currently no cure for breast cancer, despite the fact that it is one of the deadliest diseases afflicting women. Each year, the number of fatalities from breast cancer rises dramatically. The predominant form of cancer that results in mortality among women is this particular malignancy, which is of global distribution. To live a long and healthy life, any progress in the detection and treatment of cancer is essential. Therefore, maintaining the treatment aspect and patient survival level requires a high degree of accuracy in cancer prognosis. Detection of breast cancer has been facilitated by ML techniques ever since the inception of AI. This has allowed for an earlier diagnosis, hence improving the prognosis for patients. Researchers are paying close attention to ML methods because of their effectiveness; these approaches may soon have a major influence on the prediction and early diagnosis of breast cancer (BC). This article presents a method for automated BC screening that is based on ML. To facilitate an early detection of BC, numerous classification and prognostic models are developed in this research. These models rely on ML techniques, including gradient boosting and lung branching modelling. To test and train their models, the researchers in this work used an UCI ML Repository's BC Wisconsin (Diagnostic) dataset (BCWD). Additionally, an impact of feature selection, data balance, and data preparation methods used on the input dataset. An aim of this study is to identify a best ML algorithms for predicting and diagnosing BC using metrics including F1-score, accuracy, precision, recall, and confusion matrices. A result show that LGBM performed a best among the classifiers and had the greatest accuracy (98%). Based on the Python programming language and associated libraries, all work is completed in the Jupyter notebook environment.
Language: English
Open Access status: diamond
OpenAlex ID: W7217998526
FWCI: 1.681
Cites: 33
Cited by: 21
Related to: 10