Develop a neuro-fuzzy system with a heterogeneous structure for the problem described in Question 16. Note that first you need to redefine the problem so that the system output becomes the estimated risk of breast cancer.

Question 16

Develop a back-propagation neural network for diagnosing breast cancer based on features computed from a digitised image of a fine needle aspirate (FNA) of a breast mass. The data set contains 150 cases. Each case is diagnosed as malignant or benign based on 10 real-valued features: radius, texture, perimeter, area, smoothness, compactness, concavity, concave points, symmetry and fractal dimension. The data set is located on the book’s website: http://www.booksites.net/negnevitsky.

Explain the difficulties and limitations of your diagnostic neural network.