Information Technology Reference
In-Depth Information
⎧
⎨
[
Date
:
st ring
,
2
/
Jan
/
98
,
31
/
Dec
/
02
]
,
[
Symbol
:
st ring
,
aaa
,
zzzz
]
,
[
Open
:
float
,
0
,
500
]
,
S
=
[
Close
:
float
,
0
,
500
]
,
⎩
[
High
:
float
,
0
,
500
]
,
[
Low
:
float
,
0
,
500
]
,
[
Volume
:
integer
,
0
,
310000000
]
.
Symbol
is the stock name.
Open
and
Close
are the opening and closing prices for a
stockonagivenday.
High
and
Low
are the highest and lowest prices for the stock
on that day.
Volume
is the total amount of trade in the stock on that day.
We generated subscriptions by using six template subscriptions with different
probabilities. The six templates are
T
1
=
{
(
Symbol
=
P
1
)
∧
(
P
2
≤
Open
≤
P
3
)
}
with probability 20 percent,
T
2
=
{
(
Symbol
=
P
1
)
∧
(
Low
≤
P
2
)
}
with probabil-
ity 25 percent,
T
3
=
{
(
Symbol
=
P
1
)
∧
(
High
≥
P
2
)
}
with probability 30 per-
cent,
T
4
=
{
(
Symbol
=
P
1
)
∧
(
Volume
≥
P
2
)
}
with probability 10 percent,
T
5
=
{
(
Volume
≥
P
1
)
}
with probability 5 percent, and
T
6
=
{
(
Date
≥
P
1
)(
50%
)
∧
(
Symbol
=
P
2
)(
50%
)
∧
(
P
3
≤
Open
≤
P
4
)(
50%
)
∧
(
P
5
≤
Close
≤
P
6
)(
50%
)
∧
(
High
≥
P
7
)(
with probability 10 percent.
In
T
6
, each predicate appears in the subscription with probability 50 percent. We
used
T
6
to generate irregular subscriptions.
Events in DMPSS were generated by seven template events with different proba-
bilities. The seven templates are
E
1
50%
)
∧
(
Low
≤
P
8
)(
50%
)
∧
(
Volume
≥
P
9
)(
50%
)
}
=
{
=
,
=
,
=
,
=
Date
P
1
Symbol
P
2
Open
P
3
Close
}
=
{
=
,
=
,
=
,
=
P
4
with probability 10 percent,
E
2
Date
P
1
Symbol
P
2
High
P
3
Low
}
=
{
=
,
=
,
=
}
P
4
with probability 5 percent,
E
3
Date
P
1
Symbol
P
2
Volume
P
3
with probability 10 percent,
E
4
=
{
Date
=
P
1
,
Symbol
=
P
2
,
Open
=
P
3
,
Close
=
P
4
,
High
=
P
5
,
Low
=
P
6
}
with probability 5 percent,
E
5
=
{
Date
=
P
1
,
Symbol
=
P
2
,
Open
=
P
3
,
Close
=
P
4
,
Volume
=
P
5
}
with probability 30 percent,
E
6
=
{
Date
=
P
1
,
Symbol
=
P
2
,
High
=
P
3
,
Low
=
P
4
,
Volume
=
P
5
}
with probability 20 percent,
and
E
7
=
{
Date
=
P
1
,
Symbol
=
P
2
,
Open
=
P
3
,
Close
=
P
4
,
High
=
P
5
,
Low
=
P
6
,
with probability 20 percent.
We set the minimum support threshold
minsup
to 10%. According to the six
template subscriptions and their probabilities, we obtained 4 frequent itemsets:
{
Volume
=
P
7
}
Symbol
,
Open
}
(20%),
{
Symbol
,
Low
}
(25%),
{
Symbol
,
High
}
(30%),
and
{
(10%). The maximal matching load threshold
maxload
was
set to 5%. According to probabilities of the template subscriptions,
S
NF
would larger
than 85%, as a result, the number of virtual frequent itemsets was set to 3. In the
simulation, the three virtual frequent itemsets were
Symb
−
ol
,
Volume
}
{
Open
,
Close
}
,
{
High
,
Low
}
,
and
respectively.
There were 1024 nodes in the simulation network. The number of subscriptions
and events were 10240 and 102400, respectively.
{
Date
,
Volume
}
Search WWH ::
Custom Search