You are on page 1of 5

Discovering Knowledge in Data

Errata
Chapter 6
Page 107 (middle)
Text:
CHANGE:
Page 113 (top)
Text:
CHANGE:
Text:
CHANGE:
Page 118 (bottom)
Text:
CHANGE:
Page 121 (bottom)
Text:
CHANGE:
Text:
CHANGE:
Page 122 (bottom)
Text:
CHANGE:

and income (<= $50,000 or >$50,000)


and income (<= $30,000 or >$30,000
Table 6.5 row 1 and 9 has values that are truncated, rather than
rounded.
0.1112 to 0.1113
Table 6.5 row 9 columns P-sub-L and P-sub-R:
0.833 should be in column P-sub-L, and 0.167 should be in
column P-sub-R.
3/8(0.9183) + 5/8(0.7219) = 0.7946
3/8(0.9183) + 5/8(0.7219) = 0.7956. It follows that, 0.9544
0.7946 = 0.1588 becomes 0.9544 0.7956 = 0.1588.
In Figure 6.7 (Records 5)
(Record 5)
In Figure 6.7 Savings - Med
Savings = Med
a training set of 24,986
a training set of 25,000.

Text:
CHANGE:

them have incomes below $50,000


them have incomes of at most $50,000

Text:
CHANGE:

education-num > 0.8333


education-num >= 0.8333

Page 124 (top)


Text:
CHANGE:

both education groups, capital gains and capital loss


both education groups, capital-gain and capital-loss

Page 125 (top)


Text:

second split occurs on capital loss

CHANGE:

second split occurs on capital-loss

Text:

Most of the few (756 records) who had higher capital loss had
incomes above $50,000.
In contrast, the majority of those who had higher capital loss (756
records) also had incomes above $50,000.

CHANGE:
Text:
CHANGE:
Page 126 (top)
Text:
CHANGE:

which is made on marital status


which is made on Marital_Status
value of the same attribute as earlier, capital loss
value of the same attribute as earlier, capital-loss

Chapter 7
Page 130 (bottom)
Text:
CHANGE:

If output > 0.75


If output >= 0.75

Page 137 (middle)


Text:
CHANGE:

need to be normalized to between zero and 1


need to be normalized to values between zero and 1

Page 138 (bottom)


Text:
CHANGE:

W1 A = A (1) = 0.1(0.00123) = 0.000123


W0 A = A (1) = 0.1(0.00123)(1) = 0.000123

Page 143 (top)


Text:
CHANGE:

24,986 cases
25,000 cases

Text:
CHANGE:
Page 144
Text:

less than $50,000


less than or equal to $50,000
Since over 75% of the subjects have incomes below $50,000,
simply predicted less than $50,000

CHANGE:

Since over 75% of the subjects have incomes less than or equal to
$50,000, simply predicting less than or equal to $50,000

Text:

best predictor of whether a person has income less than $50,000

CHANGE:

best predictor of whether a person has income less than or equal


to $50,000

Chapter 8
Page 155 (bottom)
Text:
CHANGE:
Page 156 (top)
Text:

Text:
CHANGE:
Page 157 (top)
Text:

+1.412 = 7.88
+ 1.412 = 7.86
In Table 8.3, row h, column Cluster Membership should be
C1
2.63
= 0.3338
7.88
2.63
= 0.3346
=
7.86

In Table 8.4, row h, column Cluster Membership should be


C1

Text:
CHANGE:

+ 0.79 2 +1.06 2 = 6.25


+ 0.79 2 + 1.06 2 = 6.23

Text:

CHANGE:

Text:
CHANGE:

2.93
= 0.4688
6.25
2.93
= 0.4703
=
6.23

is larger than the previous 0.3338


is larger than the previous 0.3346

Chapter 9
Version dated: Final Book
Page 165 (bottom)

Text:
CHANGE:

in the ac-componying box


in the ac-companying box

Page 168 (top)


Text:
CHANGE:

Node1 : (0.9 0.8) 2 +(0.8 0.1) 2 = 0.71


Node1 : (0.85 0.8) 2 + (0.8 0.1) 2 = 0.78

Page 168 (bottom)


Text:
CHANGE:

Node1 : (0.9 0.2) 2 +(0.8 0.9) 2 = 0.71


Node1 : (0.85 0.2) 2 + (0.8 0.9) 2 = 0.66

Text:
CHANGE:
Page 169 (bottom)
Text:
CHANGE:

Node 2 : (0.9 0.2) 2 +(0.2 0.9) 2 = 0.99


Node 2 : (0.85 0.2) 2 +(0.15 0.9) 2 = 0.99

Node1 : (0.9 0.1) 2 + (0.8 0.1) 2 =1.06


Node1 : (0.85 0.1) 2 +(0.8 0.1) 2 =1.03

Text:
CHANGE:

Node 2 : (0.9 0.1) 2 +(0.2 0.1) 2 = 0.81


Node2 : (0.85 0.1) 2 + (0.15 0.1) 2 = 0.75

Text:
CHANGE:

Node3 : (0.1 0.1) 2 +(0.8 0.1) 2 = 0.70


Node3 : (0.15 0.1) 2 + (0.85 0.1) 2 = 0.75

Page 172 (middle)


Text:
CHANGE:

similar than clusters than are farther apart.


similar than clusters that are farther apart.

Chapter 10
Page 188 (bottom)
Text:
CHANGE:
Page 189 (bottom)
Text:

of the antecedent alone rather of than the antecedent


of the antecedent alone rather than the antecedent

with confidence 76.9%.

CHANGE:

with confidence 76.6%.

Page 192 (bottom)


Text:

= 0.001637

CHANGE:

= 0.001636

Page 193 (bottom)


Text:
CHANGE:
Page 197 (middle)
Text:
CHANGE:
Chapter 11

from the adult database using


from the adult data set using

association rule from Table 10.3:


association rule from Figure 10.6:
The Churn data set has 25,000 records.

Page 208 (top)


Text:
CHANGE:
Page 211 (top)
Text:
CHANGE:
Page 212 (top)
Text:
CHANGE:

of loan defaults, average over all loans


of loan defaults, averaged over all loans

model 2 is preferable, provided slightly higher lift.


model 2 is preferable, by providing slightly higher lift.

Figures 6.5, 6.7, and 7.9, show that the


Figures 6.8, 6.9, and 7.10, show that the

You might also like