Upload
eshwayne
View
203
Download
3
Embed Size (px)
DESCRIPTION
Citation preview
2013/01/03 P:1/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Morphology analysis for technology roadmapping: application of text mining�� �� �� � �� � �� � � � �� �� ��
!" # $ %& '( )* + ,- . /
text mining
01 23 ,4 & 56 . 7 /8 9 56 :; <=> ? @ AB C . @ AB CD E FG H IJ KL MN
1 OP R&D ManagementR&D Management (I.F.: 2.5072.507) Volume 38, Issue 1, pages 51 Q 68, January 2008RR RRSS SS TT TT
Byungun Yoon Professor at Dongguk University (
UV W < XY Z [)
Rob Phaal is Principal Research Associate in Centre of Technology Management, Cambridge U.David Probert is Reader in Technology Management, Head of CTM Manufacturing Engineering Tripos (MET), Centre of Technology Management, Cambridge U.
Paper Information
2013/01/03 P:2/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �
\] ^_` ab c ^d` ]e fg hi jk
lm c no _` b ] a
pq rs g tu vw xy
(
rs z{ | }~ �� | � �� �)�` _� �` �` o �� �m �� a� � �
� � �� � �� �i rs z{ ��
(
tu | �� �� | �� �� �
)� � ^ �` a` �` o �
�� tu � �� i ¡¢ £
(
¤¥ � � | rs � � | ¦§
mapping)\ � �b � ^_ m ^d` ]� ¨© ª« �¬ i ¬ jk
®` ] c �b � d` ] �
¯° ±² | ³´ µ¶ g ·¸ « ¹ �º
» ¼½ ¾
Paper & Report Structure
2013/01/03 P:3/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Technology roadmapping (TRM)
• Extensively applied strategic planning, linking technology and product developments to linking technology and product developments to business goals and market opportunities in a visual frameworkbusiness goals and market opportunities in a visual framework, supporting research and capital investment decisions, and enabling communication, alignment and consensus.
• Nevertheless, many companies, especially medium and small companies, still have difficulties in implementing and sustaining roadmapping, due to a number of factors (the time, cost and effort associated with maintaining what can be a complex process).
¿À ÁÂÃ Ä )Å ÆÇ È; É KL MN & Ê ËÌ ÍÎ Ï
• Another factor that hinders the adoption of the method is a lack of quality input data on markets, competitors and technology, which often relies on knowledge gathered from expert participants in workshops.
ÐÑ 2 À ÒÓ Ô Õ Ö× Ø& Ù K ) ,- D ÚÛ
The objective of this paper is to Propose a New RoadPropose a New Road--mapping Methodologymapping Methodology• uses a systems-based technique, rather than expert-based approach
/ ÜÝ Þ ßà á 'â
• using information that has been accumulated in databases from both companies and governments.
/ã Ø ,- äå æ çè éê Ä D ÁÂ & Ã Ä
• traditional roadmaps tend to concentrate on high-level strategic planning of products and technology at a discrete point in time, the new approach aims to develop a detailed view of possible product and technology configurations that is easy to update.
; < ë Ëì & Þ í îÁÎ ÏD !" = KL @ A(
ïÝ D KL MN ð ð ñò ë Ë D
)
Introduction
J ó ôõ ö÷ ø ù . úû ü
1.
/ '( ,- & !" # $ Øý J þ ÿ ,- � �
2.
( / * +� � : O� 8 9 ,- Ô (
� � 23 + :
)
� à 23 ,4
3./ * +� �D E F . :> ? !"
(
# $& 23 Å Æ
)
@ AB C
4.@ AB C H� / : �� !" & KL D � . > � �� � :D �� Þ � & MN
2013/01/03 P:4/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �
\] ^_` ab c ^d` ]� fg hi jk
lm c no _` b ] a
pq rs g tu vw xy
(
rs z{ | }~ �� | � �� �)�` _� �` �` o �� �m �� a� � �
� � �� � �� �i rs z{ ��
(
tu | �� �� | �� �� �
)� � ^ �` a` �` o �
�� tu � �� i ¡¢ £
(
¤¥ � � | rs � � | ¦§
mapping)\ � �b � ^_ m ^d` ]� ¨© ª« �¬ i ¬ jk
®` ] c �b � d` ] �
¯° ±² | ³´ µ¶ g ·¸ « ¹ �º
» ¼½ ¾
Paper & Report Structure
2013/01/03 P:5/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Technology roadmapping (TRM) /� P � �� � � !"
/
KL
VS
ÒÓ �� � D H� !
"# $ % &' () *+ � , -. / , 0 12 3
Background
4 5 67 8 http://en.wikipedia.org/wiki/Technology_roadmap
2013/01/03 P:6/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Knowledge discovery in textual databases (Text Mining) � ð Ç ÈB CD * + ,- 9: . � à 1 H ; � < =D ,4
Background
4 5 67 8 > ?@
Data Mining
A B CD E
FG H
/
I JH KL M J
NO A B FGH
/
I J H KL P Q RS T
U I A B P VI FG H
/
WXY RZ [
U I A B\ WX] ^ Z [
=>
_` ab ] ^
c O def Jg h
2013/01/03 P:7/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Morphology analysis
"# $ % &' () *+ � , -. /i 0 1j k
Background
2013/01/03 P:8/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �
\] ^_` ab c ^d` ]� fg hi jk
lm c no _` b ] a
pq rs g tu vw xy
(
rs z{ | }~ �� | � �� �)�` _� �` �` o �� �m �� a� � �
� � �� � �� �i rs z{ ��
(
tu | �� �� | �� �� �
)� � ^ �` a` �` o �
�� tu � �� i ¡¢ £
(
¤¥ � � | rs � � | ¦§
mapping)\ � �b � ^_ m ^d` ]� ¨© ª« �¬ i ¬ jk
®` ] c �b � d` ] �
¯° ±² | ³´ µ¶ g ·¸ « ¹ �º
» ¼½ ¾
Paper & Report Structure
2013/01/03 P:9/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �ConceptMN H / � lm n op q B
In this paper, roadmaps are classified using two dimensions
ObjectiveP rs turs tu ø vD # $ . w ñ� x� x � :D H� ��
level of applicationP y / z{ ñ Y âY â
//
Y |Y | Ô ! }~ ? }! }~ ? } Ô �� ~ !"�� ~ !"
Morphology-based TRM
ú* D ��
2013/01/03 P:10/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Concept�� �� �� � 3� �� �� �
Morphology-based TRM
2013/01/03 P:11/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Process &' �� � �
(
�� � ��
)
The overall process consists of three modules, each of which includes sub-processes.
Morphology-based TRM
2013/01/03 P:12/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Applications 3� �� � ��
Market Pull vs Technology Push�� ÒÓ � � & KL � �
Radical vs Incremental innovation�� ò ë Ë D � � ß� Ï& � �ì � � ¡ ¢
Advantages •
� Ü£ Ô ; � B C . ¤¥ �¦ ��
. •
ö÷ KL MN H� íÀ § ¨& Î Ï
•
© 2 KL * ª« ¬ à . ®� * ª 9: � à 1 ,4 ¤¥ !" ¯ �° ±
.
Disadvantages•
I È < =² ³ . ÿ; o @ A´ µ& � � 23 +> ? ¶·
. •
¸¹ ò ºD 23 +& »¼ D @ A´ µ½ ¾¿ !1 H ÀÁ ò ÂD MN
Morphology-based TRM
2013/01/03 P:13/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �
\] ^_` ab c ^d` ]� fg hi jk
lm c no _` b ] a
pq rs g tu vw xy
(
rs z{ | }~ �� | � �� �)�` _� �` �` o �� �m �� a� � �
� � �� � �� �i rs z{ ��
(
tu | �� �� | �� �� �
)� � ^ �` a` �` o �
�� tu � �� i ¡¢ £g ÃÄ
(
¤¥ � � | rs � � | ¦§
mapping)\ � �b � ^_ m ^d` ]� ¨© ª« �¬ i ¬ jk
®` ] c �b � d` ] �
¯° ±² | ³´ µ¶ g ·¸ « ¹ �º
» ¼½ ¾
Paper & Report Structure
2013/01/03 P:14/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Module 1: product analysis �� � �
Step 1
/ 2Å D ÆÇ � & ÈÉ Ê :; o !" @ A´ µ
morphological matrix is constructed from the attributes and levels that a subject can be characterized
Step 2 !" # $ %: . � �& � à 1 © 2 23 +
keywords are extracted from product documents that are collected for the specified product area. Text mining is applied to execute this process
Step 3
/ 23 + 1 �� !" D # $ @ A
identify the configurations of existing products, starting by matching extracted keywords with particular levels in the morphological matrix
Step 4
01 Ï !" D H� ��
compose the list of new product opportunities
Methodology
Figure 2. Example of the conversion of keyword vector into morphological matrix.
2013/01/03 P:15/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Module 2: technology analysis �� � �
Step 1
� Ë = 01 v Ç !" # $è é D © 2 KL . & KL Â D ÌÍ 2 ëdefine technology trees for the product explored in the first module
Step 2 � � ÆÇ KL É Ê& ; < KL @ A´ µ
morphological matrix of the technology is defined in the same way as for the first module. However, analysts must determine for which level in the technology tree the morphological matrix will be made
Step 3
/ * +� � 01 KL * ª: D 23 +
extract keywords from technology documents by text mining
Step 4 1 Î� KL D @ A . 23 + = Ï A´ µ! ×
identifies the configurations of existing technologies, matching keywords with the morphology matrix
Methodology
2013/01/03 P:16/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Module 3: mapping �� � �� Ð ÑÒ %Ó ÔÕ Ö
Step 1 Ã × 2 ëì B CØ !" = KL @ A´ µ ë EÙ :
link the two morphology matrices by analyzing the correlation between each level.
This linkage is based on the co-occurrence analysis of keywords
Figure 3. Example of morphology-related keywords vector.
Methodology
2013/01/03 P:17/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Module 3: mapping �� � �� Ð Ñ # �� Ú Û $ % ÜÝ Õ Ö
Step 2 Ã! × B C ; <
(
!" = KL
)
D © 2 ´ µ . Þ ß ÆÇ !" à ì © 2 D KL & É Ê
As a result of the pair-wise analysis, a correlation matrix is constructed
Examine technological requirements to meet product needs by referring to the correlation matrix
Figure 4. Example of conversion process through correlation matrix.
Methodology
2013/01/03 P:18/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Module 3: mapping �� � �� Ð ÑÒ %Ó ÔÕ Ö
Step 3
Ø KL �� á â N ã� . ä&
module1
D !" MN å Ê
visualize the plan of technology development, aligned with the product roadmap that has already established in the first module.
Methodology
2013/01/03 P:19/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �
\] ^_` ab c ^d` ]� fg hi jk
lm c no _` b ] a
pq rs g tu vw xy
(
rs z{ | }~ �� | � �� �)�` _� �` �` o �� �m �� a� � �
� � �� � �� �i rs z{ ��
(
tu | �� �� | �� �� �
)� � ^ �` a` �` o �
�� tu � �� i ¡¢ £
(
¤¥ � � | rs � � | ¦§
mapping)\ � �b � ^_ m ^d` ]� ¨© ª« �¬ i ¬ jk
®` ] c �b � d` ] �
¯° ±² | ³´ µ¶ g ·¸ « ¹ �º
» ¼½ ¾
Paper & Report Structure
2013/01/03 P:20/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Data collection&' æç èé ê
• the product manuals of Nokia mobile phones and patent documents regarding specific technologies of mobile phones were collected from internet websites. In total, 8484 product manuals were collected, covering the period 1995 Q 2004,
•• 77 77 patent documents related to the antenna technology for mobile phones filed during the same period, extracted from the US Patent and Trade Organizations (USPTO) database.&' � ë Ú ìí î � ï ðñ òó � ôõ
• The collected data were separated into two sets to develop (train) and test the proposed approach.
• Training data associated with the period 1995 Q 2002 were used to visualize past trends of products or technology
• test data from the period 2003 Q 2004 were analyzed to validate the product Q technology roadmaps produced on the basis of the training data.
Illustration
2013/01/03 P:21/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Product analysis
•
� � 'â ö÷ . ½ ? øù ú !"!" @ A´ µ rs J
7
Z à ì
/
23 � . û n à ì �
ÆB õ ã ü � ýÉ Ê
•
þÿ * ª: �
461
n © 2 D 23 + . � È: � �
48
n 23 + = !" @ A´ µD à ì &
É Ê� 2�
•
� ý à ì D ò � É Ê . �õ
648
ýò � D @ A
648 (i.e. 3 � 3 � 3 � 2 � 3 � 2 � 2=648)
Illustration
2013/01/03 P:22/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Product analysis
•
9 Á ÒÓ
(1995-2002)
� � D
53
ý Ø � !" . H� rs �
31
ý @ A•
� :
(2003-2004)
ÏD !" �� & H� @ A �
648 Q 31=617
ý
•
� l 1
2001-2004
Ò p Ø �D !" @ A
Illustration
� P1.
8617
ý H� @ A w ÿ �à 'â & ÒÓ �� ��� ß ¶· � � ° � ñ� ¯ �
2.
32
ý !" � � 9: r
s 1
19
ý @ A � > ÷ �
�� D H� ì
3.
8
19
ý H� @ A: . �� ý � : � ÒÓ �1
¦ . È�
13
ý� Î�D !" � � � � : .8 � � H� / � ð !"D �õ @ A : �� � :!" D @ A& ��
(
 ��� ö÷ Þ ÈD H ?ì
)
2013/01/03 P:23/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Technology analysis
•
/ KL !
(Technology tree)
"¥ > ? KL B C . rs ÆÇ KL Â D þ ÌÍ•
½ KL @ A´ µ rs J
6
Z à ì
/
23 � . û n à ì Æ� ã ü � ýÉ Ê• 77
� '( * ª ��
226
n 23 + . � È: � �
56
n = KL @ A � 2• 1995-2002
D
69
� '( . H� rs 1
61
ý @ A
•
9 Á Ò p �D
53
ý Ø � !" . � ��
31
ý KL @ A � �
Illustration
Figure 5. Technology tree of mobile phone.
2013/01/03 P:24/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Technology analysis
•
'( KL @ A � �6 # À !" @ A D � �6
•
8 ñ÷ n �� D �¦ . Å J '( $ ÿ à % $D Þ ß . &' ñ � × ¦ � D '( ( )�
•
8 ÷ E F á *P û n '( ½ +� È, D @ A . ò � À È- '( �
Illustration
2013/01/03 P:25/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Product .. .. Technology roadmap
•
/ m n @ A´ µ ÆÉ Ê × yD © 2 0 6 . 1� !" @ A = KL @ A  2 0�•
*
7
� � P û n2 D !" @ A × yD KL @ A& 2
(
� � ñ 3 ZDPearson
© 2 0 6
)
Illustration
4545 4545
(
67 89 :)
•
úû üD ; !" ; = ; KL ; <J= ë Ë D > ? 6 . © 2 ì Þ � y@ /Pearson
As Chi-Square Test
2013/01/03 P:26/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Product .. .. Technology roadmap
•
1 Ï !" @ Aè × y Ô é ÿD KL � A . 1� ñ � ( )~ H> ? �� �
Illustration
� P1.
� *8
: . '( @ A
1 &3 B C ¯ � . ä �
2000
D =
2002
DB � E Y '( FùG H õ Iý �
2.
J F ÿ ¯ � Ï !"
1
K .L � 9 G M N OP æ Q ©2 '( KL R Þ ßè é ©2 KL '( . � &S ò �
�T '( ( )�
2013/01/03 P:27/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Product .. .. Technology roadmap
•
!"
-
KL U VN � � !" = KL @ A � D W X� � ú YZ D [ V KL J Z . � é
ÿ\ ý KL @ A í� ] ¨B ^_ � ý !" @ A D ÿ `�
•
� aD Ø È H� b Z y / � È� c � à ì ~ !"
Illustration
Figure 6. Product Q technology roadmap of mobile phone
2013/01/03 P:28/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Validation de fg h ij k lm nde fg h ij k lm nde fg h ij k lm nde fg h ij k lm n
•
/o pq
(1995-2002)
01 D \ ý '( KL @ A . Þ x < × rs q Â
(2003-2004)Â D '
( . E F �¦ È: m Ç '( C tu
=>
þÿ w� m ýv � 1 ¦ . � w ñ � � ó ö÷ Þ
Èw D H / ì
•
'â ö÷
In this research, key aspects of the method were reviewed by experts (designers and technologists) from LG Electronics
Illustration
x IP* y: v � ¡ ü8 9
LG
'â × & ö÷ Þ Èw D 1w . � � 7{z � 'â D 5& � ] ¡ ¢ KL ¯ �& { rD Ê &ì
) À | ó }~ D w u ÷ � �=>
8 nè � ÜÝ ì D
R&D
Þ Èw .Ç Èà á 'â D à ��(
þÿ 'â { rD Ê &ì ®ò �
)
2013/01/03 P:29/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �
\] ^_` ab c ^d` ]� fg hi jk
lm c no _` b ] a
pq rs g tu vw xy
(
rs z{ | }~ �� | � �� �)�` _� �` �` o �� �m �� a� � �
� � �� � �� �i rs z{ ��
(
tu | �� �� | �� �� �
)� � ^ �` a` �` o �
�� tu � �� i ¡¢ £
(
¤¥ � � | rs � � | ¦§
mapping)\ � �b � ^_ m ^d` ]� ¨© ª« �¬ i ¬ jk
®` ] c �b � d` ] �
¯° ±² | ³´ µ¶ g ·¸ « ¹ �º
» ¼½ ¾
Paper & Report Structure
2013/01/03 P:30/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Managerial implications�� � � � �� �� � # � �� � �� � ��
• Risk î M N = � º� . Ê & M � Ë 1 . è é ! ÉD !" = KL @ A� (à ì > 6 Ô
Æ à ì D É Ê 6
)
• Accurate use of quantitative analysis requires careful identification, selection and filtering of source data
P o� ò º �
/
23 + Ç È� & × y& »¼ � � . � ¾¿ »¼ ¶· E F&
»¼ D @ A � w
=>Text Mining
D ��
•
!" B C = KL B C H� � �> ? . � Þ í> ?m � Â Dmapping
• MA-based TRM can be practised in both a top-down (market pull) and bottom-up (technology push) orientation.
Conclusions
2013/01/03 P:31/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Limitations and Future Research�� � � � �� �� �� �
1.
@ AB C � � O� Â �D > ? 6
(Attribute)
& É Ê
(Level) . Ç È O� ë Ë ? 6 .
EX
P ÁÂ Ô � Ô ¡ ¢£ þÿ H� Ø !" 6 ; Ô ¤ $ Ô ¥ / Á � ) ë Ë ? 6¦§ J ò � É ÊD > ? 6 . �� Á w ñ� � ¨õ ¼ ©& Öª2.
ö÷ Þ Èw ò ñ / : ° � « ¬ ® . � � 01 ¯ 9 !" = KL @ A  D ! É� ��
° � . v � ± Èà á 'â ² �& � z ° ±
3.
� ³ Ò p � Hà D � ¯ ,- . 5 O ¤´ & Ê &ì ~ µò J Æâ �� D ¶ � �· *
ª�¸ æ¹ º � » � �
•
( / ~ ¯ � È� D 6 ; Þ È : H ¼½ u Æ à ì D @ A
(
äå � u ¾ ÷ ¿ � � O� >
? 6 D ��
)
•
À / à Á ¤´ ~ KL ¤´ D B C . : "¥ !" KL ¯ � , ° ±& « ¬ ®
(
äå� u ¾ã ¿
)
•
Î H ¼D :à | rs = �� 8 Ä Þ Èw . 3 �� / à | !" ¯ �� (
Å ú
)
Conclusions
2013/01/03 P:32/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �
\] ^_` ab c ^d` ]� fg hi jk
lm c no _` b ] a
pq rs g tu vw xy
(
rs z{ | }~ �� | � �� �)�` _� �` �` o �� �m �� a� � �
� � �� � �� �i rs z{ ��
(
tu | �� �� | �� �� �
)� � ^ �` a` �` o �
�� tu � �� i ¡¢ £
(
¤¥ � � | rs � � | ¦§
mapping)\ � �b � ^_ m ^d` ]� ¨© ª« �¬ i ¬ jk
®` ] c �b � d` ] �
¯° ±² | ³´ µ¶ g ·¸ « ¹ �º
» ¼½ ¾
Paper & Report Structure
2013/01/03 P:33/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Æ Ç
g È ng È ng È ng È n
Consistency9
�
(
� �
) î
•
!" # $ %& '( ,-
=>
/
text mining
01 23 ,4 & 56
=>
/ 8 9 56 :; <=> ? @ AB C
=>
@ AB CD E FG H IJ KL MN & !" KL ¯ � � � �
•
v ± È J ÈË ÌÍÎ 'â ö÷ &Ï J ² �
(
w ñÐ � ³ 'â à � = ö÷)Ñ Ò nÑ Ò nÑ Ò nÑ Ò n
Completeness 10
�
(
� �
) î Ó ÔÕ Ö ×Ø � Ù
• Morphology Analysis, Technology Road Mapping & Text mining
ÚÛ Í Ì º �Ü á
•
� ÝÞ º � Ô ! ß1 Ã | �Z �
•
àÃ Ü á � û ü ¿ á
â ã nâ ã nâ ã nâ ã n
Correctness7
�
(
� �
) î Ó äå � æ ç � Ù•
,-è B õ Þ È �� & �� m ��•
à | y / é ¿À ,- D � &ì Ô Text Mining
& @ A´ µD Ê &ì )Å Æ
• (
�
)
!" à ì ò ) À
(
� :)
ÒÓ �� & �� é `
•
'( @ A ò ) À � : KL ê � & � �
(
KL �ë . EX:
ì í Ø �à á © �&
PC)
PS
î ïð ð ñòó ôõö÷ ø ùú ûüý þ ÿ ��
2013/01/03 P:34/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Æ Ç
�� � � k ���� � � k ���� � � k ���� � � k �� �� ��
•
/
Text MiningText Mining
½ � ð Ç È O� D * + ì ,- ¦ ;ì ,- > ÷ � O�•
� Ë H�
1.
� Û �� � ~ �� D * +� : § z
=>
� <* � � D '( �� = # $ Øý
� � ; ��
2.
H �u Ï >
/
�� [â Ô �� 'â :� ¥ ¡ ¢* +� �& � Ë � �D � &ì
g �� � �� k !" h ig �� � �� k !" h ig �� � �� k !" h ig �� � �� k !" h i
(
@ AB C ò � # $Ï %&
)•
>' D (� � T L D " ) ~
TRIZ
Þ Èw
•
? 6 Â D 2 ëì . O� �ò ÷ � � * © 2 0 6 Z+ . / × y Þ C
(Correspondence Analysis, CA)
Î , ( Ô � ��(
® H� O� ë Ë ? 6
)
2013/01/03 P:35/35
� � � �� � � � �� � � � � �� �� � � � � � � � � � � �Q&A
ú* D !" = KL B C >' À T L D " ) -�
m � Â D
Fitting/Mapping . Ã . � $ �/ 0 ñ
trial and error
& 1 Õ [ 2D E F
(
� = 3 v � ]- 4~ { r
)