The problems of development of Information Control Systems with the factor of high data quantity growth are discussed. Hierarchical structure of the system goals is proposed. It is shown the role of unification and formalization of documents, which circulate in the System.
Все статьи журнала
Журнал:
In this work conditions of existing of orthogonal algebraic bases in lineals and complete orthogonal systems in (closed with respect to certain topology) subspaces of indefinite inner product spaces are investigated.
Журнал:
Let $H$ be a Hilbert space and let $A: \mathscr{D}(A) \subset H \to H$ be a selfadjoint positive definite operator, $A^{-1} \in \mathscr{G}^q(H)\ (q>0),\ \beta_l \gt 0\ (l = \overline{0, m}), \ 0 = : b_0 \lt b_1 \lt \cdots \lt b_m$.
Журнал:
In this article the notion of the template of set functions and various kinds of such templates are introduced. Are considered primary k-parametrical templates of a kind t = t[+, •].The Vapnick Chevronenkis Dimension (VCD) for image of such templates is estimated. As well as VCD for the class of polynomials above a field of real (rational) numbers and the class of Zhegalkin polynomials are estimated.
Журнал:
The solvability of the smooth transition equation in the Hilbert space double-index scale well adjusted for Fourier transform and convolution type equation solving in considered.
Compact component and abbreviated interval structures of attribute space generated by empirical data
Журнал:
The compact component structure of attributes space, clustering algorithm based on this concept, the brief interval structure concept of attributes space and its construction algorithm are considered in the paper.
Журнал:
This paper is devoted to the description of all irreducible representations of some real involutive algebras on Hilbert quaternionic bimodules, naturally arising is resulting at consideration of nonequivalent involutions of graduated analogue of Lee algebra о(2,H).
Журнал:
The problem of multicriteria Pseudo-Boolean optimization for the case of partially given objective function is considered. The analysis of algorithms (branches and borders method, criterion linear fold, locals search) for solving bicriteria problems of Pseudo-Boolean linear optimization is carried out when criteria are represented by the information on the coefficient signs.
Журнал:
The problem of finding the most appropriate subset of features or regressors is the generic challenge of Machine Learning problems like regression estimation or pattern recognition. We consider the problem of time-varying regression estimation, which implies also the inevitable necessity to choose the individual appropriate levels of model volatility in each of regressors, ranging from the full stationarity of instant models to their absolute inde-pendence in time.
Журнал:
Problems arising up at the decision of tasks of operative regional and department management are considered in the work. Management methods are offered for thousands of remote organizations on the basis of heterogeneous information.
Журнал:
Theoretical preconditions and program realization of computer system of indexing and abstracting of text documents on the basis of knowledge base use about a subject domain in the form of an is situational-syntagmatic network are considered. The approach offered by authors to calculation informative lexemes and the offers, based on comparison of statistical characteristics of word forms in the reviewed text and a full corpus is used in this system.
Журнал:
In this article the grayscale image skeletonization problem is shown by the example of images of fingertips prints. The main idea: the grayscale image skeleton is constructed using some parts of binary images skeletons. One can take these binary images after initial grayscale picture binarization with various brightness levels. Such binary images skeletons fragments give us the opportunity constructing clear binary image of initial grayscale picture.
Журнал:
A lot of image analysis problems lend themselves to a unified mathematical formulation as optimization problems, namely (min, +)labelling problems. It is known that these problems are NP-hard in the general formulation, but when the adjacency graph is a tree, they can be effectively solved by the tree-serial dynamic programming procedure. In this paper, it is considered an optimization method on the basis of Gauss-Seidel principle, which uses iterative reevaluation of groups of variables with tree-like neighborhood relation.
Журнал:
We consider systems described by initial-boundary value problems for parabolic equations with discontinuous coefficients. From observations of the state of systems, we find minimax prediction estimates
Журнал:
This paper is dedicated to the problem of image retrieval from large databases of biomedical images using the query by example paradigm. The specific characteristics of biomedical images are briefly discussed in the context of similarity retrieval and possible ways of image searching are outlined. A versatile method is suggested which is based on extended multi-sort and multi-dimensional co-occurrence matrices. The use of the method is illustrated on a variety of image retrieval tasks using real databases of biomedical images of different types.
Журнал:
The paper considers multidimensional scaling problem with proximity matrix consisting of two different non-diagonal elements. Elements of the superdiagonal matrix don’t decrease along the string and don’t increase along the column. The structure of considered matrix is described. In one-, two- and three-dimensional spaces all available proximity matrices are listed. Matrices that can be considered as proximity matrices in t-dimensional euclidian space have their dimensions in a definite interval. The exact bounds of this interval are also found.
Журнал:
The notion of the key antichain for the lattice of characteristic space interval descriptions, key antichains family property, the algorithm of the key antichains construction are conside the paper.
Журнал:
An algorithm for calculating dense disparity map for stereo images is proposed. Stereo correspondence problem is formulated in terms of an assignment problem in this algorithm. The algorithm processes each image scan-line individually similar to a dynamic programming approach to the dense stereo correspondence problem, however unlike dynamic programming, it does not rely on ordering constraint. Ordering constraint elimination is useful for the scenes with narrow objects in front of the camera. The proposed algorithm supports occlusions detection and guaranties uniquness of found matches.
Журнал:
The results received last years within the limits of the algebraic approach to the recognition problems are described. Algebraic closures of algorithms of classical estimation algorithm model are investigated.
Журнал:
The minimax estimations of the scalar product ⟨l, x(N)⟩ are obtained on the basis of observations y(k) = H(k)x(k) + η(k) up to N − 1 moment assuming that l ∈ L ⊆ Rn, x(k) is a solution of the difference equation S(k + 1)x(k + 1) = A(k)x(k) + f(k), x(0) = f0 with some rectangular matrix S(k) of rank p(k), f0, f(k) – are unknown vectors, η(k) – is random vector with unknown correlation matrix R(k), M η(k) ≡ 0.
Журнал:
The estimate for the assymetry of a face of a man by its 3D model is proposed in the paper. It contains the detailed algorithm constructed for this estimate calculation. The used face modes are obtained by a methof of 3D scanning and are represented in the form of points cloud. The problem of detection of a face symmentry plane is solved. The algorithm was tested on 200 samples of face models. This testing shows the stability of the proposed estimate.
Журнал:
Global optimization problems associated with cryptographic analysis of asymmetric ciphers. The aim of this article is to establish relation between well-known problems of cryptographic analysis and global optimization problems which can be associated with SAT representation of cryptographic algorithms where bits of key is part of SAT solution string. There was constructions SAT forms for factorization problem, SAT forms for logarithmic problem and logarithmic problem on elliptic curves.
Журнал:
The double inequality V CD(S) ≤ Kl(S) < V CD(S) log l is proven, where V CD(S) is the Vapnik —Chervonenkis Dimension of some collection of general recursive functions, Kl(S) is the Kolmogorov’s complexity of this collection S, l is the length of data set (sample) which is shattered by S. The novel pV CD approach for V CD estimation based on above inequality is suggested.
Журнал:
The automatic classification (cluster analysis) complex algorithm, that was especially developed for intellectual complex-organized data handling and decision support problems, is described.
It includes: the m-local optimization algorithm with the predetermined classification performance criterion, the informative parameters selection algorithm, the initial fragmentation algorithm, the missing observation filling algorithm.
Журнал:
The analysis and forecasting within the loosely-formalized multivariate control system, consisting of sufficiently large number of a priori non-structured objects, problem solution method is proposed. As a forecasting model for each object the markovian chain with rstates, where r - the number of structural units (classes), is used.
Журнал:
Authors describe the technique developed by them of predicate word's possible deep syntactic valences. S characteristics's values's automatic retrieval, Initial data are elements of lexicographic interpretation. The idea of algorithm of recognition of semantic valencies's role orientations is in more detail considered.
Журнал:
All groups concidered in this paper are finite. The main aim in this note is receipt bounds of derived π-length for a π-soluble group depending on nilpotent π-length and derived length of π-Hall subgroups.
Журнал:
In this paper,we consider the problems arising when input-output operations are planed. The main factors defining data processing on input-output stage are presented. The new approach to resolve above problems are presented.
Журнал:
In this paper,we consider the problems arising when input-output operations are planed. The main factors defining data processing on input-output stage are presented. The new approach to resolve above problems are presented.
Журнал:
The paper considers approach to the image recognition problem that is based on random Gibbs fields. Estimation of parameters of Gibbs models. One of the possible approach to this problem is maximum
entropy principle. According to this principle we should seek such distribution which is as non-informative as available prior information. the paper suggests another approach, which seeks such distributions, for which maximum and minimum values of probabilities of random object is attained.