Resumen

MARTINEZ-GUZMAN, Gerardo; CERON-GARNICA, Carmen; FERNANDEZ-PEREZ, Jorge Alejandro  y  VILLEGAS-CERON, Gerardo. Breast Cancer Classification through Mixture of Bivariate Normal Using EM Algorithm. Comp. y Sist. [online]. 2024, vol.28, n.4, pp.2221-2230.  Epub 25-Mar-2025. ISSN 2007-9737.  https://doi.org/10.13053/cys-28-4-5296.

An analysis is presented in this paper for benign and malignant diagnosis of tumors, biopsies have shown an increase of nuclear size, and changes in the texture of the tumor nucleus. In this article, an analysis is made using the unsupervised learning algorithm Expectation-Maximization (EM). Two variables are analyzed: the mean of the radius and texture of the tumors, being the former a measure of the average distances from the center of the tumor to its perimeter, and the latest is the variance of gray-scale values. Since the behavior of the said variables is similar to the mixture of normals in two opposing categories. The EM algorithms demonstrates ability to categorize the dataset into two different labels (malignant and benign). This model projects a classification with a high percentage of coincidence with the observed data.

Palabras llave : Maximum likelihood estimators; breast cancer; EM algorithm; Gaussian mixture model.

        · texto en Inglés     · Inglés ( pdf )