Ávila Hernández, Ricardo, Rossell Mendoza, Kevin Ricardo, Soto Mora, Josue Alejandro (Universidad La Salle México, Facultad de Negocios, 2020)
Machine Learning comprises a wide range of models aimed at solving real life problems using supervised and unsupervised algorithms capable of finding even the finest causalities and correlations between any given phenomena portrayed in data. Given the current extraordinary software capabilities, we can exploit this tool in practically any field – Oncology. For instance, a medical speciality which focuses on Cancer treatment can make use of these models to provide a more accurate diagnosis when it comes to Breast Cancer Detection. In this article we delve into a catalogue of Machine Learning models and discuss their effectiveness through specific criteria in order to choose the most suitable one for this problem. The Analytic Hierarchy Process displayed conclusive results assigning to the Random Forest the highest scores in each one of the analyses employed, over 10% better than the Logistic Regression, the second highest evaluated model in the overall analysis. The models we developed with data describing different features of different breast tumour nuclei, therefore, for another type of data results may differ.
