Skip to main content

Table 1 Prior studies about the weight optimization methods

From: A rank weighted classification for plasma proteomic profiles based on case-based reasoning

Authors

Year

Methods

Weights

Cardie & Howe [14]

(1997)

Information gain

G(f) a

Ahn & Kim [12]

(2009)

Relative importance [0-7]

\( \frac{x_f}{\sum \limits_{f=1}^m{x}_f} \)

Gu et al. [13]

(2010)

Delphi method

–

Chang et al. [5]

(2011)

Delphi method

–

Zhao et al. [7]

(2011)

Entropy method

\( \frac{entropy_f}{\sum \limits_{f=1}^m{entropy}_f} \)

Liang et al. [8]

(2012)

Logistic regression

\( \frac{Wald_f}{\sum \limits_{f=1}^m{Wald}_f} \)

  1. a indicates information gain of the f-th feature, and entropy is defined as \( -\sum \limits_i{p}_i\bullet {\log}_2{p}_i \)