Database Reference
In-Depth Information
4. clothes F 3=true clothes F 5=true 127
→
clothes F 4=true 88
conf:(0.69)
<
lift:(10.62)
>
lev:(0) [79] conv:(2.97)
5.
clothes F 2=true
414
→
clothes F 5=true
101
conf:(0.24)
<
lift:(9.83)
>
lev:(0.01) [90] conv:(1.29)
...
14.
clothes F 4=true
1089
→
clothes F 1=true
91
conf:(0.08)
<
lift:(5.79)
>
lev:(0) [75] conv:(1.07)
15. lingerie 2=true 267
→
clothes F 4=true 99 conf:(0.37)
<
lift:(5.68)
>
lev:(0) [81] conv:(1.48)
16. clothes F 4=true 1089
→
lingerie 2=true 99 conf:(0.09)
<
lift:(5.68)
>
lev:(0) [81] conv:(1.08)
17. household 1=true clothes F 3=true 162
→
clothes knitting=true 83
conf:(0.51)
<
lift:(5.2)
>
lev:(0) [67] conv:(1.83)
18. clothes knitting=true 1645
→
household 1=true clothes F 3=true 83
conf:(0.05)
<
lift:(5.2)
>
lev:(0) [67] conv:(1.04)
19. household 1=true clothes F 4=true 209
→
clothes knitting=true 103
conf:(0.49)
<
lift:(5)
>
lev:(0) [82] conv:(1.76)
20. clothes knitting=true 1645
→
household 1=true clothes F 4=true
103 conf:(0.06)
<
lift:(5)
>
lev:(0) [82] conv:(1.05)
Again, clothes stores are abundant in the rules. The top 14 rules even exclu-
sively contain clothes stores selling women's fashion. The rest of the top 20 is
completed with rules that also link with a household store, a lingerie store, and
a clothes store that also sells knitting accessories. Clearly, this shows that strong
associations exist between stores that are focused on a more female-oriented
public. It might be interesting to focus on clothes stores that sell men's fashion
(clothes M) in order to zoom in on a male audience. When we filter out the rules
that do contain such a store in their itemset, we end up with the following 4
rules:
49. clothes M 1=true 385
→
snacks=true 83 conf:(0.22)
<
lift:(4.12)
>
lev:(0) [62] conv:(1.2)
50. snacks=true 874
→
clothes M 1=true 83 conf:(0.09)
<
lift:(4.12)
>
lev:(0) [62] conv:(1.08)
197. clothes M 1=true 385
→
household 3=true 111 conf:(0.29)
<
lift:(2.08)
>
lev:(0) [57] conv:(1.21)
198. household 3=true 2318
→
clothes M 1=true 111 conf:(0.05)
<
lift:(2.08)
>
lev:(0) [57] conv:(1.03)
We find associations between one men's clothing store and a snacks store and
household store respectively. Although these rules are clearly less strong (low
confidences), it is noteworthy that other and less trivial associations are found