We describe numerical and analytical investigations of causal sets sprinkled into spacetime manifolds. The first part the paper is a study finite Alexandrov subsets Minkowski dimensions $1 + 1$, 2$ 3$. In particular we consider rank 2 past causet events, which set events that are two links to past. Assigning one as `preferred past' for each event yields structure', was recently proposed basis d...

This paper introduces novel methods for feature selec­ tion (FS) based on support vector machines (SVM). The methods combine feature subsets produced by a variant of SVM-RFE, a popular feature ranking/selection algorithm based on SVM. Two combination strategies are proposed: union of features occurring frequently, and ensemble of classifiers built on single feature subsets. The resulting method...

This paper introduces novel methods for feature selec­ tion (FS) based on support vector machines (SVM). The methods combine feature subsets produced by a variant of SVM-RFE, a popular feature ranking/selection algorithm based on SVM. Two combination strategies are proposed: union of features occurring frequently, and ensemble of classifiers built on single feature subsets. The resulting method...

This paper introduces novel methods for feature selec­ tion (FS) based on support vector machines (SVM). The methods combine feature subsets produced by a variant of SVM-RFE, a popular feature ranking/selection algorithm based on SVM. Two combination strategies are proposed: union of features occurring frequently, and ensemble of classifiers built on single feature subsets. The resulting method...

In the theory of classical matroids, there are several known equivalent axiomatic systems that define a matroid, which described as matroid cryptomorphisms . A q -matroid is -analogue where subspaces play role subsets in theory. this article we establish -matroids. doing so highlight difference between and its -analogue. We introduce comprehensive set axiom show them existing -matroid. These ax...

This paper introduces novel methods for feature selec­ tion (FS) based on support vector machines (SVM). The methods combine feature subsets produced by a variant of SVM-RFE, a popular feature ranking/selection algorithm based on SVM. Two combination strategies are proposed: union of features occurring frequently, and ensemble of classifiers built on single feature subsets. The resulting method...

This paper introduces novel methods for feature selec­ tion (FS) based on support vector machines (SVM). The methods combine feature subsets produced by a variant of SVM-RFE, a popular feature ranking/selection algorithm based on SVM. Two combination strategies are proposed: union of features occurring frequently, and ensemble of classifiers built on single feature subsets. The resulting method...

We prove that the filling order is quadratic for a large class of solvable groups and asymptotically quadratic for all Q-rank one lattices in semisimple groups of R-rank at least 3. As a byproduct of auxiliary results we give a shorter proof of the theorem on the nondistorsion of horospheres providing also an estimate of a nondistorsion