Information Technology Reference
In-Depth Information
Actor1's Model of Actor2
10
9
8
Music1(8)
7
Music2(3)
6
5
Music3(7)
4
3
Music4(5)
2
1
0
0
100
200
300
400
Runs
Fig. 14.4
How Actor 1's model of Actor 2's evaluative scale changes over 300 runs
Table 14.1
Actor 1's range of beliefs of actor 2 after 300 runs
(model)
x /
Music
0
1
2
3
4
5
6
7
8
9 Indifference
I(x)
Expected
E(x)
Series 1
0.01 0.01 0.02 0.03 0.04 0.07
0.1
0.15
0.23 0.3
0.16
7.18
Series 2
0.5
0.24 0.12 0.06 0.03 0.02
0.01
0.01
0.01
0
0.24
1.09
Series 3
0.04 0.07
0.1
0.12 0.14 0.14
0.13
0.12
0.09 0.1
0.11
4.72
Series 4
0.09 0.13 0.15 0.15 0.14 0.12
0.09
0.06
0.04
0
0.11
3.52
0-9 such that actor 1 has some expectation as to each scale position concerning a
piece of music for each actor. The sum of these beliefs for a series adds up to one since
the actor must place the music somewhere on the scale. There are three important
measures derived from the ranges of belief:
Log
2
−
1
x
p
x
.
log
2
p
x
1
.
Indifference, I
(
x
)
=
−
Where p
x
is the belief of x for a given actor's perception of another actor's view of
a piece of music. I(x) represents the value that a belief would need to have if, for
the same level of overall uncertainty (entropy), the belief value were to be equal for
all hypotheses (scale positions). Under this hypothetical situation all the hypotheses
(scale values in this case) would be indifferent to each other. We take this level of
indifference to be a threshold above which the hypotheses are considered 'believed'
and below which they are not. This threshold is dynamic and tends to become higher
as more hypotheses' belief values approach zero.
x
p
x
.
x
2
.
Expectation E(x)
=
