Theoretical and Empirical Analysis of ReliefF and RReliefF
Citations
1,971 citations
Cites background from "Theoretical and Empirical Analysis ..."
...One algorithm, from individual evaluation, is ReliefF (Robnik-Sikonja and Kononenko, 2003) which searches for nearest neighbors of instances of different classes and weights features according to how well they differentiate instances of different classes....
[...]
...We use three synthetic data sets to illustrate the strengthes and limitations of FCBF and compare it with ReliefF, CFS-SF, and FOCUS-SF....
[...]
...Comparison between FCBF(0) and ReliefF shows that ReliefF is unexpectedly slow even though its time complexity is linear to dimensionality....
[...]
...For each data set, we conduct Student’s paired two-tailed t-Test in order to evaluate the statistical significance of the difference between two averaged accuracy values: on resulted from FCBF(log) and the other resulted from one of FCBF(0), the full set, ReliefF, CFS-SF, and FOCUS-SF....
[...]
...For ReliefF, we use 5 neighbors and 30 instances throughout the experiments as suggested by Robnik-Siko ja and Kononenko (2003)....
[...]
1,566 citations
Cites background from "Theoretical and Empirical Analysis ..."
...Some representative criteria include feature discriminative ability to separate samples (Kira and Rendell 1992; Robnik-Šikonja and Kononenko 2003; Yang et al. 2011; Du et al. 2013; Tang et al. 2014), feature correlation (Koller and Sahami 1995; Guyon and Elisseeff 2003), mutual information (Yu and Liu 2003; Peng et al. 2005; Nguyen et al. 2014; Shishkin et al. 2016; Gao et al. 2016), feature ability to preserve data manifold structure (He et al. 2005; Zhao and Liu 2007; Gu et al. 2011b; Jiang and Ren 2011), and feature ability to reconstruct the original data (Masaeli et al. 2010; Farahat et al. 2011; Li et al. 2017a)....
[...]
...ReliefF (Robnik-Šikonja and Kononenko 2003) selects features to separate instances from different classes....
[...]
...Some representative criteria include feature discriminative ability to separate samples (Kira and Rendell 1992; Robnik-Šikonja and Kononenko 2003; Yang et al. 2011; Du et al. 2013; Tang et al. 2014), feature correlation (Koller and Sahami 1995; Guyon and Elisseeff 2003), mutual information (Yu and…...
[...]
1,353 citations
1,056 citations
857 citations
Cites methods from "Theoretical and Empirical Analysis ..."
...We show that two powerful feature selection algorithms, ReliefF ( Robnik-Sikonja & Kononenko, 2003 ) and Laplacian Score (He et al., 2005) are special cases of the proposed framework....
[...]
...We show that two powerful feature selection algorithms, ReliefF (Robnik-Sikonja & Kononenko, 2003) and Laplacian Score (He et al., 2005) are special cases of the proposed framework....
[...]
...Supervised feature selection algorithm ReliefF ( Robnik-Sikonja & Kononenko, 2003 ) is a special case of SPEC by setting b ’(¢) = b ’1(¢), ∞(L) = L and deflning W as:...
[...]
...Supervised feature selection algorithm ReliefF (Robnik-Sikonja & Kononenko, 2003) is a special case of SPEC by setting .f(·) = .f1(·), ....
[...]
References
21,694 citations
21,674 citations
"Theoretical and Empirical Analysis ..." refers background or methods in this paper
...5 (Quinlan, 1993)) and for regression it is the mean squared error (MSE) of average prediction value (used in e....
[...]
..., 1984) or Gain ratio (Quinlan, 1993) in classification and mean squared error (Breiman et al....
[...]
17,177 citations
14,825 citations