A neurofuzzy route to breast cancer diagnosis and treatment

Abstract

In this paper an outline is given of a modelling approach, using neurofuzzy networks, to be used in an intelligent oncology workstation for the improved treatment and diagnosis of breast cancer. This neurofuzzy approach is intended to assist in the provision of the most suitable treatment and therapy for the individual patient in this important medical domain and to seek to add to knowledge in this vital area to yield improved diagnostic and treatment techniques. The major component of the system is a high-dimensional approximator neurofuzzy network (Adaptive Spline Modelling of Observation Data or AS-MOD) which is a constructive learning algorithm used to automatically generate high-dimensional approximations and to identify complex relationships between input variables and the measured output to form models which may be interpreted as sets of linguistic fuzzy rules.<>