218 /302
分析台灣金控公司之關鍵風險因子：以風險平衡計分卡結合決策分析實驗室法為基礎之分析網路法
218
(1)
(2)
The ANP, a generalization of the AHP, can consider interrelations among factors
(Saaty, 2001). The waterfront development surveys in the studies conducted by Lee, Wu,
Hu, and Flynn (2013) and Lee, Wu, Suthiwartnarueput, Hu, and Rodjanapradied (2016)
have shown that the results of the ANP were more meaningful than those of the AHP.
However, the lengthy and complex pairwise comparisons in ANP questionnaires are time
consuming and are a burden on questionnaire respondents. To enhance the feasibility and
effectiveness of the survey process, the DANP transforms the total influence matrix of the
DEMATEL approach into an unweighted supermatrix of the ANP. Following the
DEMATEL procedures, the total influence matrix was assessed according to direct
influence scores inputted by respondents, which express the strength of the influence of
the binary relationship between every two variables, as illustrated in Table 4. For the RM–
BSC, the respondents must fill a 4 × 4 matrix and a 21 × 21 matrix for the dimensions and
criteria, respectively.
Table 4 Illustrative Input Direct-Influence Matrix
→
A
B
C
D
A
1
3
0
B
0**
2
3
C
4
3
1
D
0
4*
3
Note: 0. No influence; 1. Low influence; 2. Medium influence;
3. High influence; 4. very high influence.
*Criterion D has a very high influence on criterion B considering its importance.
**Criterion B has no influence on criterion A considering its importance.
More applications and details of the DEMATEL and DANP are provided in the
studies by Wu (2008), Yang, Shieh, Leu, and Tzeng (2008), Tsai and Chou (2009), Chen,
Hsu, and Tzeng (2011), and Chiu, Tzeng, and Li (2013). In this study, we applied the
standard DANP steps developed by Chen et al. (2011) and Chiu et al. (2013):