Humaninteractive annealing for multilateral observation

1
Human-interactive annealing for multilateral
observation

• Yoshiharu Maeno, Kenichi Horie, Kiichi Ito, and
  Yukio Ohsawa

2
New chance discovery
Objective discovery of relevant events neither
visible nor observed as data

• Application example
• Discovery of an emerging technology in RD
• Discovery of a leader hidden in an organization
• Discovery of unconscious preference for marketing

3

• Be aware of weak relationships in observation
  beyond human's intuition
• Imagine their unobserved origin.

Something shall be discovered there.

• Humans interpretation
• Crystallization algorithm

Interactive annealing
Invisible relevant events Hypothetical scenarios
4
Three advantages of the interactive annealing
process

• Fusion of humans work and data processing
• Heritage from chance discovery
• Extension to multilateral observation

• Humans interpretation
• Adaptive nature to the individual humans degree
  of understanding

• Crystallization algorithm
• Stable and deterministic nature

5
Interactive annealing process
6
Heating step
7
Cooling step
8
Crystallization algorithm
Step 1 Item identification
Basket
Basket
Basket
Basket
a,b,c
h,I,j,k,l
d,e,f,g
a,b,d,h
Step 2 Clustering
Step 3 Dummy item insertion
Basket
Basket
Basket
Basket
a,b,d,h,D4
a,b,c,D1
h,I,j,k,l,D3
d,e,f,g,D2
D1,D2,D3,D4
Step 4 Co-occurrence calculation
Step 5 Topology analysis
9
Extension to multilateral observation
Another observation is referred to enrich humans
interpretation.
10
Example for US patent analysis (1)
11
Example for US patent analysis (2)
12
Three advantages of the interactive annealing
process

• Fusion of humans work and data processing
• Heritage from chance discovery
• Extension to multilateral observation

• Humans interpretation
• Adaptive nature to the individual humans degree
  of understanding

• Crystallization algorithm
• Stable and deterministic nature