show the steps of how you arrived at the answers Decision Tree Suppose you need to learn a decision tree from a set of d
Posted: Fri May 20, 2022 3:06 pm
show the steps of how you arrived at the answers
Decision Tree
Suppose you need to learn a decision tree from a set of data
with a population of 13, and there are two classes (assume 6 are
positive and 7 are negative). Now suppose a test on attribute A
breaks the population into two groups - one group with a mix of
2(+) and 7(-), and the other group with 4(+). If we test on
attribute B first, it then breaks the population into two groups
with 6(+) and 1(-) in one group and 6(-) in the other.
Calculate the initial entropy, then state which attribute (A or
B) should be tested first, by showing their respective information
gains.
Hint: if done by a scientific calculator, note
log2(x) = log10(x)/log10(2) =
ln(x)/ln(2)
Decision Tree
Suppose you need to learn a decision tree from a set of data
with a population of 13, and there are two classes (assume 6 are
positive and 7 are negative). Now suppose a test on attribute A
breaks the population into two groups - one group with a mix of
2(+) and 7(-), and the other group with 4(+). If we test on
attribute B first, it then breaks the population into two groups
with 6(+) and 1(-) in one group and 6(-) in the other.
Calculate the initial entropy, then state which attribute (A or
B) should be tested first, by showing their respective information
gains.
Hint: if done by a scientific calculator, note
log2(x) = log10(x)/log10(2) =
ln(x)/ln(2)