How is the coefficient of variation used in the Delphi method for weight optimization?

Conditions

• The coefficient of variation $C$ must be less than 0.25 for the judgment to be valid.
• Experts provide pairwise comparisons of factor importance.

Reasoning, step by step

1. Calculate the coefficient of variation $C = \frac{\sqrt{\frac{1}{n}\sum_{i=1}^{n}(x_i - \bar{x})^2}}{\bar{x}}$ using the experts' judgments.
2. Check if $C < 0.25$.
3. If valid, average the judgments to determine the final weight vector $W$.

Example

The video displays the formula for $C$ and states that if $C < 0.25$, the judgment is considered valid.

Common misconceptions

• Assuming that a higher coefficient of variation indicates better consensus among experts.
• Forgetting that the threshold 0.25 is a specific condition for validity in this context.

Watch the explanation

Explore next

• What are the factor set and comment set in fuzzy comprehensive evaluation?
• How does the paper say the final label excellent is obtained from the displayed vectors?