A comparative study of linear discriminant analysis and an artificial neural network performances in breast cancer diagnosis

The use of different artificial intelligence techniques is increasingly widespread around the world. Among the various possibilities it offers, one that has been gained prominence in recent years was the medical field whether in diagnoses or treatments. In this scenario, using the Wisconsin Breast Cancer Database's parameters of a Fine Needle Aspiration, this paper proposes a performance comparison between an artificial neural network and linear discriminant analysis on a breast cancer diagnosis system, since this disease is responsible for thousands of deaths and its effective diagnosis is fundamental in increasing the chances of cure. Therefore, in this article are made analysis of the direct results of these applications and alternatives to improve the performance. Finally, it was obtained a 95.77% rate of correct diagnoses for the linear discriminant analysis and 92.78% for the artificial neural network.