2b) I do not understand this question at all. Do I need to list all possible splits or do i need to do some calculation here? for example,

• Do we just list down like university education we can split by YES or NO
• Income we can split by
  • Low vs (medium and high) (Low and medium) vs high
• Profession we can split by
  • white vs (blue and self)
  • self vs (blue and white)
  • blue vs (white and self)

2. The table below shows the records from a country club membership application database. University Education Income Level No High Yes Medium No Medium Yes High Yes Low Profession Self-employed White-collar White-collar Blue-collar Self-employed Status Rejected Accepted Rejected Accepted Rejected Status is the class label attribute. We are building a binary classification tree using GINI index to measure impurity of the data. (a) What is the GINI index for the observations in this data set? (b) List all possible splits that must be considered at the root node. (c) Which of these two attributes: (University Education, Income Level] will not be selected as the attribute for splitting the root node? Explain your answer. (Note: the attribute that gives lower information gain should not be selected)