Therefore, it is desirable to construct an attribution model to help us
in visualizing these associations and even expand the analysis beyond the initial core property
collection to ensure coverage of the problem domain that is being addressed.
How can an attribution model be utilized to further analyze discovered core attributes
and later on derive concepts? Why does the business concept extraction require a more distilled
Attribute A Attribute B
Core Attributes
Attributes
A+B
Attributes
A+C
Attributes
B+C
A+B+C
Level 1
Level 2
Level 3
Converging Level
Attribute C
Requirements
EXHIBIT 4.4 ATTRIBUTION MODEL: A NEURAL NETWORK
Establishing an Attribution Model 79
Return
Time
Return
Liquidity
Time
Liquidity
Return
Risk
Time
Risk
Liquidity
Risk
Return
Time
Liquidity
Return
Time
Risk
Time
Liquidity
Risk
Return
Risk
Liquidity
Return
Time
Liquidity
Risk
Level 1 ??“ Core Attributes
Level 2
Level 3
Level 4 ??“ Converging Level
Return? Time? Liquidity? Risk?
Requirements
EXHIBIT 4.5 TRADING ATTRIBUTION MODEL EXAMPLE
process? An attribution model is analogous to a network grid that is composed of various nodes,
each of which represents one or more attributes, as depicted in Exhibit 4.4. In essence, this
concept is founded on the neural network formation,1 a popular computer model that resembles
human brain structures and information transmission and learning activities.
Pages:
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155