hmk2solstud1

# Hmk2solstud1 - Joseph Scarabino HW 2 pt 1 Problem 1 1 2 Evaluate Predictive Accuracy Tree 1 3 well classified 6 total = 50 Tree 2 2 well classified

This preview shows pages 1–7. Sign up to view the full content.

Joseph Scarabino 10/13/11 HW 2 : pt 1 Problem 1 1)

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
2)

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Evaluate Predictive Accuracy: Tree 1: 3 well classified / 6 total = 50% Tree 2: 2 well classified / 6 total = 33% Problem 2 Tree 1: Obj Age Income Student Credit Buys_comp 1 >40 High No Fair Yes 2 <=30 Low Yes Excellent Yes 3 31-40 Low No Excellent No 4 <=30 High No Fair Yes 5 >40 Medium No Excellent No 6 <=30 Medium Yes Fair Yes Tree 2: Obj Age Income Student Credit Buys_comp 1 >40 Low Yes Fair Yes 2 31-40 High No Fair Yes 3 <=30 High No Excellent No 4 31-40 Medium No Excellent Yes 5 >40 Low Yes Excellent Yes <=30 Medium Yes Fair No Problem 3 Predictive Accuracy from notes: 4 well classified / 6 total = 67%
Joseph Scarabino 10/13/11 106310863 HW 2 : pt 2 Problem 1

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
I will calculate Information Gain on the two nodes chosen after the initial Student node (age & credit). I(p,n) = -p/(p+n) * log 2 ( p/(p+n) ) - n/(p+n) * log 2 ( n/(p+n) ) E = (p i + n i )/(p + n) * I(p i ,n i ) Gain(A) = I(p,n) – E(A) Age: I(3,4) = -3/(3+4) * log
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 01/25/2012 for the course CSE 352 taught by Professor Wasilewska,a during the Fall '08 term at SUNY Stony Brook.

### Page1 / 8

Hmk2solstud1 - Joseph Scarabino HW 2 pt 1 Problem 1 1 2 Evaluate Predictive Accuracy Tree 1 3 well classified 6 total = 50 Tree 2 2 well classified

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online