Database Reference
In-Depth Information
5.6 Experimental Results
We conduct a systematic empirical study using a real data set and some synthetic
data sets on a PC with a 3.0 GHz Pentium 4 CPU, 1
0 GB main memory, and a 160
GB hard disk, running the Microsoft Windows XP Professional Edition operating
system. Our algorithms were implemented in Microsoft Visual C++ V6.0.
.
Tuple
R
1
R
2
R
3
R
4
R
5
R
6
R
7
R
8
R
9
R
10
R
11
R
14
R
18
Drifted days 435.8 341.7 335.7 323.9 284.7 266.8 259.5 240.4 233.6 233.3 232.6 230.9 229.3
Membership
0.8
0.8
0.8
0.6
0.8
0.8
0.4
0.15
0.8
0.7
0.8
0.6
0.8
probability
Top-10 prob.
0.8
0.8
0.8
0.6
0.8
0.8
0.4
0.15
0.8
0.7
0.79
0.52 0.359
Table 5.2
Some tuples in the IIP Iceberg Sightings Database 2006.
Rank 1 2 3 4 5 6 7 8 9 10
Tuple
R
1
R
2
R
3
R
5
R
6
R
9
R
9
R
11
R
11
R
18
Pr
(
t
,
j
)
0.8 0.64 0.512 0.348 0.328 0.258 0.224 0.234 0.158 0.163
Table 5.3
The answers to the
U-KRanks
query.
5.6.1 Results on IIP Iceberg Database
We use the International Ice Patrol (IIP) Iceberg Sightings Database
1
to examine
the effectiveness of top-
k
queries on uncertain data in real applications. The Inter-
national Ice Patrol (IIP) Iceberg Sightings Database collects information on iceberg
activities in the North Atlantic. The mission is to monitor iceberg danger near the
Grand Banks of Newfoundland by sighting icebergs (primarily through airborne
Coast Guard reconnaissance missions and information from radar and satellites),
plotting and predicting iceberg drift, and broadcasting all known ice to prevent ice-
bergs threatening.
In the database, each sighting record contains the sighting date, sighting location
(latitude and longitude), number of days drifted, etc. Among them, the number of
days drifted is derived from the computational model of the IIP, which is crucial in
1
http://nsidc.org/data/g00807.html
