Build a decision tree by calculating the overall entropy and the entropy for each independent variable. Which attribute would be used for the first split only (no need to build the whole decision tree)? Show all work. Your answer may be kept in fraction form. (10 points)

The following table shows some software metrics (independent variables) and whether the software has a high number of defects, or low number of defects.

Independent Variables			Dependent Variable
Program Vocabulary	Complexity	Lines of Code	Software Defects
Moderate	High	Few	Few
Moderate	High	Many	Many
Low	High	Many	Few
Moderate	High	Many	Few
Moderate	Low	Many	Few
Moderate	Low	Few	Many
Moderate	Low	Many	Many
High	Low	Many	Many