Multi-Layer Combinatorial Fusion Using Cognitive Diversity

The DIRAC framework: Geometric structure underlies roles of and in combining classifiers

“…We are actively pursuing these and additional expansions of the DIRAC framework (Sniatynski et al., manuscript in preparation and work in progress). Likewise, this provides one theoretical anchor to ground work examining how to integrate the information in scores into the DIRAC framework, potentially linking our other work in rank-score diversity 12 , 44 and cognitive diversity 40 , 45 with our work in DIRAC (here and Sniatynski et al. 19 ).…”

“…When these are extended to the Euclidean space, the symmetric groups feature regular geometric structure which forms a convex polytope of dimension N − 1, embedded as a manifold within the containing N-dimensional space (e.g., B 3 is a 2-dimensional hexagon embedded in a three-dimensional space). 32 , 33 , 36 , 37 , 39 , 40 For example, Figures 3 A–3C show the three permutation polytopes (also referred to as permutahedra) that may be readily displayed within three dimensions; these correspond to the bubble sort Cayley graphs containing two, three, and four elements denoted B 2 , B 3 , and B 4 , respectively. These three permutation polytopes are the line, the hexagon, and the truncated octahedron.…”

Improving SDG Classification Precision Using Combinatorial Fusion

“…The combined models perform consistently as well as, and in many cases outperform, the best of the two individual models. Other relevant publications include those discussing the rank space [18,32] and those in the field of metric fixed point theory that can be useful to further this methodology [33,34].…”

“…Combinatorial fusion algorithm (CFA) provides methods and algorithms for combining multiple scoring systems using the rank-score characteristic (RSC) function and cognitive diversity (CD) [3,4]. It has been used widely in protein structure prediction [5], ChIP-seq peak detection [6], virtual screening and drug discovery [7,8], target tracking [9], stress detection [10,11], portfolio management [12], visual cognition [13], wireless network handoff detection [14], combining classifiers with diversity and accuracy [15], and text categorization [16], to name just a few (see [17][18][19] and the references within).…”

“…by sorting the score values using the rank value as the key [3,4,18]. The notion of the RSC function was proposed by Hsu, Shapiro, and Taksa [4] in information retrieval.…”