Information Technology Reference
In-Depth Information
terms from relevant category. Authors of CTD claim that incorporating only rele-
vant feature can be highly effective and perform comparatively well with other
measures, especially on collection with highly overlapped topics [9]. Since the lei-
sure entertainment event could be unfolded in several categories for comprehensi-
ble description. We utilize CTD method as alternative comparative method for
providing reference information for recommendation. Since the decision quality is
subjective to user's perception and their social context, the original performance
measuring matrix is replaced by cognition parameters in this chapter.
The proposed CTD method is extend from TF-IDF, where TF refers to term
frequency in category
c
and ICF is interpreted as inverse category frequency. TF-
ICF scheme shows no way of discriminating between terms that occur frequently
in a small subset of documents and terms that are present in a large number of
documents throughout a category. The formula of CTD was defined as follows.
CTD
(
t
,
c
)
=
TF
(
t
,
c
)
⋅
IDF
(
t
,
c
)
⋅
ICF
(
t
)
k
k
i
k
i
k
where
⎛
C
⎞
⎜
⎜
⎝
⎟
⎟
⎠
ICF
(
t
)
=
log
k
CF
(
t
)
k
⎛
D
(
c
)
⎞
⎜
⎜
i
⎟
⎟
⎠
IDF
(
t
,
c
)
=
log
k
i
DF
(
t
,
c
)
⎝
k
i
D( c
i
)
denote the number of document in category
c
i
C
denote number of category in the collection.
CF(t
k
)
denote the category frequency for term
t
k
DF(t
k
,c
i
)
denote the document frequency for term
t
k
in category
c
i
The CTD method also utilizing the tags from heterogeneous information sources
for tag classification and weighting. For the purpose of taking the classification is-
sue into consideration, CTD method is also used in this chapter for contribute
proximal social intelligence for leisure recommendatory service.
4 i-Bike Leisure Recommendatory Service
Based on the concept of proximity from social network relationships, we propose
a collaborative leisure entertainment recommendatory service, called “i-Bike”.
The i-Bike service explores those proximal social intelligence from both context
and content and enable quality decision making. The i-Bike service illustrates
and exchanges user's personal experience for bicycle exercise entertainment in
Taiwan. A schematic diagram of interactions is shown as Figure 1. The process
can be unfolded into two parts, one is the experience contribution process, and the
other is knowledge acquisition process.
