Information Technology Reference
In-Depth Information
type-related features
article-related features
(a)
(b)
40
40
30
30
20
20
10
10
0
0
0.000
0.002
0.004
0.006
0.008
0.000
0.005
0.010
0.015
0.020
0.025
0.030
bits
bits
customer-related features
feedback-related features
(c)
(d)
50
60
50
40
40
30
30
20
20
10
10
0
0
0.000
0.001
0.002
0.003
0.004
0.005
0.1
0.2
0.3
0.4
0.5
0.6
bits
bits
Fig. 8.11
Distributions of mutual information for the four types of features,
a
type-related,
b
descriptive,
c
customer-related,
d
feedback-related
Table 8.1
Features with
highest mutual information
excluding feedback-related
features
Feature
Type
Mutual Information
Care instructions
Descriptive
0.0132
Article type class 5
Type
0.0083
Brand
Descriptive
0.0067
Article type class 4
Type
0.0059
Article type class 3
Type
0.0050
Jacket preference
Customer
0.0047
Shoe preference
Customer
0.0043
Age
Customer
0.0042
Jeans preference
Customer
0.0042
Weight
Customer
0.0041
information completely. Excluding the feedback-related attributes, we detail the next
best features in Table
8.1
.
We observe little discriminative power even by the most informative descrip-
tive, type-, and customer-related features. Care instructions and brands carry most
information in the scope of descriptive features. The different type classes stick to
Search WWH ::
Custom Search