Risk Detection of Malignant Tumors in Mammograms using Unconventional Computing

In this paper, we propose the use of Alpha-Beta associative approach as an Unconventional Computing method in the pre-diagnosis of malignant tumors of breast cancer, obtaining an accurate result in a simple way; trying to avoid invasive diagnostic methods like biopsies, as far as possible. This proposal provides for the Alpha-Beta Support Vector Associative Machine created in 2008 and tested for classification of binary images. The results show that the classification model to detect malignancy is very competitive compared to others of the best known classification methods, having an accuracy of 81.85%.