表 3.

基于多准则决策的权重分配算法伪代码

Pseudocode of the proposed multi-criteria decision-making-based weight allocation algorithm

Algorithm 1 Pseudocode of the Proposed Method
Input:
An evaluation matrix D ∈ Rn × m, where n is the number of evaluation objects, represent for multi-classifier or multi-modality, m is the number of evaluation criteria.
Process:
Step 1: Normalize the evaluation matrix D column by column by:
Step 2: Allocating weight for each evaluation object i by:
;
Step 3: Define the weight uniformization factor K :
;
Step 4: Uniformize the weights by the defined weight uniformization factor K to make sure the sum of weights equals to 1:
.
Output:
Weight for each evaluation object ωi(i = 1, 2, ..., n).