l. Suppose you are required to classify a tumour a.s benign or malignant on the basis of 9 features of the tumour, such as e.g. uniformity of cell size, clump thickness, mitosis, etc. You have a data set of 699 ca.se records of tumour. For each ca.se record of a tumour, you have the 9 features describing the tumour, together with correct classification of this tumour a.s benign or malignant.

(a) Design a neural network model for the classification task.

(b) If the performance of your model on the training set is good, but the test set performance is significantly worse, what is the most likely problem? Discuss a simple way which may improve your results.

(c) If you need even more accurate results, what kind of approaches would you try?