Problem
We wish to perform classification of Breast Cancer test samples from a source data set. The dataset consists of 569 samples with 30 feature measurements per sample. Your job is train on 10% of the data to see how well it will classify the remaining 90% of the data using a Support Vector Machine approach. The data points are to be classified into 'Benign' and 'Malignant' based on the data available.
i. Perform PCA on the data set to reduce the data set to 2 dimensions from 30. Plot the points for the two classes from your training set.
ii. Perform Fisher Linear Discriminant (FLD) analysis to reduce the dataset to 2 dimensions. Plot the points for the two classes from your training set.
The dataset is taken from: K. P. Bennett and O. L. Mangasarian: "Robust Linear Programming Discrimination of Two Linearly Inseparable Sets", Optimization Methods and Soft- ware 1, 1992, 23-34.