Table 3 Well known DR methods from the literature that were compared with the proposed Vec2im-Siam
DR approach | Acronym | Description |
|---|---|---|
Infinite Latent Feature Selection (Roffo et al. 2017) | ILFS | A probabilistic latent feature selection approach that performs the ranking step by considering all the possible subsets of features bypassing the combinatorial problem |
Unsupervised graph-based filter (Roffo et al. 2015) | Inf-FS | In Inf-FS, each feature is a node in the graph, a path is a selection of features, and the higher the centrality score, the most important the feature. It assigns a score of importance to each feature by taking into account all the possible feature subsets as paths on a graph. |
Relief-F (Liu and Motoda 2007) | Relief-F | An iterative, randomized, and supervised approach that estimates the quality of features according to how well their values differentiate data samples that are near to each other; it does not discriminate among redundant features, and performance decreases with few data. |
Laplacian Score (He et al. 2005) | LS | The importance of a feature is evaluated by its power of locality preserving. In order to model the local geometric structure, this method constructs a nearest neighbor graph. LS algorithm seeks those features that respect this graph structure. |
Fisher filter feature selection (Gu et al. 2011, Xue-qin et al. 2006) | Fisher | It computes a score for a feature as the ratio of interclass separation and intraclass variance, where features are evaluated independently, and the final feature selection occurs by aggregating the m top ranked ones. |
Correlation-based Feature Selection (Shahbaz et al. 2016) | CFS | CFS sorts features according to pairwise correlations |
Unsupervised Feature Selection with Ordinal Locality (Guo et al. 2017) | UFSOL | A clustering-based approach that preserves the relative neighborhood proximities and contributes to distance-based clustering |
Least Absolute Shrinkage and Selection Operator (Hagos et al. 2017) | Lasso | This method applies a regularization process that penalizes the coefficients of the regression variables while setting the less relevant to zero to respect the constraint on the sum. FS is a consequence of this process when all the variables that still have non-zero coefficients are selected to be part of the model |
Chi-square feature selection (Thaseen and Kumar 2017; Thaseen et al. 2018) | Chi2 | It ranks features based on the statistical significance test and consider only those features that are dependent on the class label |
Minimum redundancy maximum relevance (Nguyen et al. 2010) | mRMR | A FS algorithm that systematically performs variable selection, achieving a reasonable trade-off between relevance and redundancy. |
Fuzzy Complementarity Criterion (Moustakidis et al. 2012, Moustakidis and Theocharis 2010) | FuzCoC | FS is driven by a fuzzy complementary criterion which assures that features are iteratively introduced, providing the maximum additional contribution with regard to the information content given by the previously selected features |