Breast tumor classification using a new OWA operator
2016; Elsevier BV; Volume: 61; Linguagem: Inglês
10.1016/j.eswa.2016.05.037
ISSN1873-6793
AutoresEmad A. Mohammed, Christopher Naugler, Behrouz H. Far,
Tópico(s)Artificial Intelligence in Healthcare
ResumoBreast cancer is the most common cancer among Canadian women and the second cause of death from cancer. Fine needle aspirate (FNA) is a technology used to investigate early breast tumors to detect cancer. In this paper, we demonstrate the application of a new ordered weighted averaging operator (OWA) to the problem of breast tumor classification. The OWA operator employs the Laplace distribution to calculate the weight vector to aggregate the uncertain information about the breast tumors. The aggregated information is used along with the tumor label, i.e., benign or malignant, to train a nearest neighbor, support vector machine, and logistic regression classifiers. The result of this study based on the nearest neighbor classifier achieves 99.71% accuracy that outperforms other studies that utilize other OWA operators using the same dataset.
Referência(s)