predictive analytics; classification

Question 1

qualitative outcome

Answer 1

binary variable, yes or no, represented by numerical values 1 and 0

Question 2

regression does not work for classification because

Answer 2

betas can push predicted outcomes past its bounds (0-1)

Question 3

dummary/binary variable

Answer 3

way of transforming categorical variable in numerical values, and can often be binary variable

Question 4

dependent and independent variables in classification

Answer 4

dependent = binary
independent = can be continuous
eg y = fruit, x = height, width

Question 5

to read a classification tree

Answer 5

outcome on top node is overall prediction
- 0.smth under this represents probability that the outcome (throughout tree) is whatever was decided by the tree
- percentage under this represent proportion of data that fits into ctieria

Question 6

accuracy rate

Answer 6

correct predictions/number of observations