Abstract: Shot boundary detection is main step in video management systems like browsing and indexing. In this paper, we shortly describe an earlier proposed shot detection algorithm based on the structural properties of video frames. Two mathematical models for decision making, i.e. for similarity threshold determination are proposed and compared. The first model allows determination of threshold in case of using mean-square deviation for frames similarity and when the image pixel is assumed to be normal distributed random variable. The second model is based on using the structural similarity measure based on Weibull model for image gradient magnitude distribution. Global and adaptive approaches are considered for frames similarity threshold determination. Results of experiments to detect real video shots are given. It is shown that adaptive threshold determination method generally gives more acceptable results than global threshold determination method. At the same time the 2 W based threshold determination approach gives more accurate results than PSNR based approach, therewith corresponds to HVS perception.

Keywords: video segmentation, shot detection, similarity measure, threshold determination

ACM Classification Keywords: Image Processing and Computer Vision

Link:

METHOD FOR VIDEO SHOT DETECTION AND SEPARATION

David Asatryan, Manuk Zakaryan

http://www.foibg.com/ijima/vol03/ijima03-03-p04.pdf

Abstract: Method of domain models designing is proposed in this paper. Domain models are represented as class diagrams. Initial information for domain models designing contains both analytical representation of domain data streams and domain entities. Domain data streams are defined using collaboration diagram. Analytical representation of domain data streams as a subset of Cartesian product of the following sets: objects and messages are proposed. Also an approach of decomposition collaboration diagram into fragments for defining data streams is represented. Rules of defining relations between classes of domain model by means of analysis both analytical and graphical representation of domain data streams are formulated. Both description of domain entities and interconnections between them are represented in terms of algebra describing software static models. Proposed method involves an iterative approach for domain model designing. Firstly it is necessary to obtain information about class diagram constituents. Then information about methods of classes is complimented using analytical representation of domain data streams. Using both analytical representation of domain data streams and analytical description of entities relations between class diagram constituents are defined and clarified. An example of designing domain model for domain “designing cutting schemas for leather goods details” is represented. Also information about interconnections of domain entities and domain processes is represented.

Keywords: class diagram; collaboration diagram; model transformation; model driven architecture; set-theory tool; transformation rules; cutting schemas designing.

ACM Classification Keywords: D.2.2 Design Tools and Techniques; D.2.11 Software Architectures

Link:

METHOD OF DOMAIN MODELS DESIGNING

Elena Chebanyuk

http://www.foibg.com/ijima/vol03/ijima03-03-p03.pdf

Abstract: Context: The use of software repositories is recent in software engineering. Different techniques have been applied to software engineering problems. We want to know the degree of application of neural networks as data mining technique. The results have allowed the preliminary state of the art of the application of neural network techniques on software repositories.

Keywords: neural networks, repository, data mining.

ACM Classification Keywords: I.2.6 Artificial Intelligence –Connectionism and neural nets, H.2.7 Database Administration – Data warehouse and repository.

Link:

A STUDY OF APPLICATION OF NEURAL NETWORK TECHNIQUE ON SOFTWARE REPOSITORIES

Ana Maria Bautista, Tomas San Feliu

http://www.foibg.com/ijima/vol03/ijima03-03-p02.pdf

Abstract: We review the development of the infinity concept and continuous mathematics from ancient times to the present day. To the critical analysis of the application of the infinitesimal in the mathematical modeling in XVIIXVIII centuries is given special attention. It is noted necessity of development methods for the analysis of finite small disturbances in computer simulations. The problem of the adequacy of the linear approximation of mathematical and computer models is under discussion. We investigate the influence of finite small disturbances into elements of model to quality of localization solutions of the model which is based on the methodology of options sequential analysis. The results of numerical experiments and taken into account the level of systems conditionality and algorithms (the mantissa length in the numbers representation) are given.

Keywords: continuous and infinite, infinitesimal analysis, mathematical and computer simulation, linear approximation, localization, successive analysis variants, analysis finitely small values

ACM Classification Keywords: H.4.2 Information Systems Applications: Types of Systems: Decision Support.

Link:

ANALYSIS FINITELY SMALL VALUES IN COMPUTER SIMULATION

Oleksii Voloshyn, Vladimir Kudin

http://www.foibg.com/ijima/vol03/ijima03-03-p01.pdf

Annotation: Analysis of physiological properties of the nerve fibers, associated with their specialized function – excitation transmission, is done. Possibility in principle of the same principles' usage in the communication channels is established. It is shown that characteristics of nerve fibers as communication channels are comparable to characteristics of existing wire transmission media of information in telecommunications. Prospectivity of research of excitations transmission principles as the basis for implementation of communication channels is established.

Keywords: nerve fibers, medullated (myelinic) nerve fiber, communication channel, conductor, propagation velocity.

Keywords classification of ACM: B.4.1 Data Communications Devices, B.4.2 Input/Output Devices, B.4.3 Interconnections (Subsystems), H.1.2 User/Machine Systems

Link:

ABOUT USAGE POSSIBILITY OF PRINCIPLES OF EXCITATIONS TRANSMISSION BY NERVE FIBERS FOR TELECOMMUNICATIONS

Galina Gayvoronska, Maxim Solomitsky

http://www.foibg.com/ijima/vol03/ijima03-02-p11.pdf

Abstract: A method of monitoring the state of the radio channel using statistical information of the turbo decoder, the use of which in the adaptive system will allow carrying out its restructuring under varying random actions.

Keywords: information technologies, a priori uncertainty, wireless networks, turbo codes.

ACM Classification Keywords: H.4 Information system applications, H.4.3 Communications Applications, G.1.6 Optimization, I.6 Simulation and modeling, K.6.4 System management

Link:

METHOD FOR CONTROLLING STATE CHANNEL WIRELESS NETWORKS UNDER A PRIORI UNCERTAINTY

Sergey Zaitsev

http://www.foibg.com/ijima/vol03/ijima03-02-p10.pdf

Abstract: The article is devoted to application of optical devices based on tensosensitive properties of liquid crystals for gathering physical data of biological objects in bioinformatics researches. Graphical results of prototype’s test are presented and usage recommendations done.

Keywords: biological researches, measuring devices, fiber optics, liquid crystals.

ACM Classification Keywords: J.3 Life and Medical Sciences – Bioinformatics; C.4 Performance of Systems - Measurement techniques

Link:

APPLICATION OF FIBER-OPTICAL MODULATORS AS MEASURING DEVICES IN BIOINFORMATICS RESEARCHES

Oleksandr Ryabtsov

http://www.foibg.com/ijima/vol03/ijima03-02-p09.pdf

Abstract: In this paper, an algorithm is developed for a full non-conflict schedule obtained through the activation of diagonals of sub- matrices in switching nodes of type Crossbar. The size of sub matrices in the Crossbar is optimized. Finite automats are designed for two layers control of sub-matrices a comparison with known algorithms with diagonal activation of connection matrix is done.

Keywords: Network nodes, Crossbar switch, Conflict elimination, Packet messages, finite automat.

ACM Classification Keywords: C.2.1 Network Architecture and Design, C.4 Performance of Systems

Link:

HIERARCHICAL TWO LAYERS CONTROL COMMUTATOR FOR IMPLEMENTATION OF FULLY NON-CONFLICT SCHEDULE

Kiril Kolchakov, Vladimir Monov

http://www.foibg.com/ijima/vol03/ijima03-02-p08.pdf

Abstract: Relationships and communication between doctors and patients is a very important problem in medicine. Analysis of possible interrelated directions of development and application of communicators in emergency and family medicine was performed. According to the first direction the IT communicators are intended for supporting alternative communication with the people lost temporarily or permanently possibility to speak. According to the second direction proposed communicators have databases included detailed information about types of diseases and traumas, and by results of analysis of injured person symptoms or feelings the software subsystem generates some variants of possible diseases or traumas diagnosis. It is proposed to apply special hardware-software means, particularly IT communicators to simplify the communication of doctors with patients (particularly with voice limitation) and increase treatment efficiency beginning from the first contact. Proposed IT communicators are used in family and emergency medicine for supporting first contact of doctors with patients (with voice limitation) and getting information about patient state, for example, during the preliminary examination of patients. Application of proposed modern computer devices and information technologies in medicine makes it possible to reduce a number of cognitive problems, increase the efficiency of the first contact of patients with doctors, quickly give a diagnosis and correctly select methods of treatment.

Keywords: IT communicators, problems of cognitivism in medicine, family medicine, speciality applicationdependent software.

ACM Classification Keywords: J.3 Life and Medical Sciences - Health

Link:

IT-COMMUNICATORS FOR MEDICINE

Ozar Mintser, Volodymyr Romanov, Igor Galelyuka, Oleksandr Voronenko

http://www.foibg.com/ijima/vol03/ijima03-02-p07.pdf

Abstract: Informational models of the advanced systems of radioaccess to the telecommunications networks comparison. LTE (Long Term Evolution) technology compared with the 802.1x protocol family.

Keywords: wireless networks, WPAN, WMAN, LTE, HSDPA, OFDM.

ACM Classification Keywords: C.2 Computer-Communication? Networks; C.2.1 Network Architecture and Design – Wireless communication.

Link:

INFORMATIONAL MODELS OF THE ADVANCED SYSTEMS OF RADIOACCESS TO THE TELECOMMUNICATIONS NETWORKS COMPARISON

Sergey Mikhailov, Sergey Shreyner

http://www.foibg.com/ijima/vol03/ijima03-02-p06.pdf

Abstract: A method of behavioral software models synchronization is represented in this paper. Implementing this method behavioral software models, which are changed after communication with customer, are synchronized with other software models that are represented as UML diagrams. Method of behavioral software artifacts synchronization makes the Model-Driven? Development (MDD) approach more effective. For synchronization of different behavioral software models, transformation approach in the area of Model-Driven? Architecture (MDA) is proposed. Synchronization operation is executed using analytical representation of initial and resulting models. Initial behavioral software model is represented by UML Use Case Diagram. Resulting behavioral software model is represented as UML Collaboration Diagram. Analytical representation of UML Use Case diagram allows considering data flows. For this representation set-theory tool operations are used. As a Collaboration Diagram usually contains more information in comparison with Use Case one, method defines two types of Use Case diagram fragments. From the beginning Use Case diagram fragments that can be transformed directly to resulting diagram constituents are considered. Then the rest of Use Case diagram fragments are processed to represents rules of placement Collaboration Diagram messages. These rules help to designate data flows, represented in Collaboration Diagram, more accuracy. Method, proposed in this article, can be used both separately and be a part of more complex transformation technics, methods and frameworks solving different tasks in MDA sphere. Also the example of proposed method realization for solving task “designing of Vector hodograph of density lying function” (VHDLF) is represented. A process of designing Collaboration Diagram, considering Use Case diagram and ontology knowledge analysis is represented. The constituents of Collaboration Diagram, which are designed using different sources, namely Use Case diagram, ontology knowledge and requirements specification are defined.

Keywords: Software model transformation, Use Case diagram, Set-theory tool, behavioral software model synchronization, Vector hodograph of density laying function.

ACM Classification Keywords: D.2.2. Design Tools and Techniques, D.2.11. Software Architectures

Link:

METHOD OF BEHAVIORAL SOFTWARE MODELS SYNCHRONIZATION

Elena Chebanyuk

http://www.foibg.com/ijima/vol03/ijima03-02-p05.pdf

Abstract: One of the goals of authors of the Business Rules Approach, besides making the process of specification of software more precise and adequate, was to provide a rigorous basis for reverse engineering of business rules from existing software systems. By different reasons this goal was no more a subject of interest of the involved researchers. One of the general problems of using business rules (BRs) in the reverse engineering of software is the selection of language to present the rules, extracted from the source code. In principle, it is not impossible to use for this purpose the languages dedicated for the straightforward task – using BRs for specification of software systems. The current paper presents an overview of different such languages in order to decide whether they are appropriate for rules extraction and to outline those features which will be helpful in language comparison and selection of a specific language for BR.

Keywords: software development, software modernization, software reengineering, business rules, languages for specification of business rules, automated business rules extraction, comparative analysis.

ACM Classification Keywords: D.2.1 Requirements/Specifications, D.2.7 Distribution, Maintenance, and Enhancement.

Link:

COMPARATIVE ANALYSIS OF THE LANGUAGES FOR BUSINESS RULES SPECIFICATION

Krassimir Manev, Neli Maneva

http://www.foibg.com/ijima/vol03/ijima03-02-p04.pdf

Abstract: In this paper we introduce three formulas for calculation of the weights of the MWHD through the notion of a degree of friendship/relationship Γ. Further investigations and applications of this decision making procedures are also proposed, which are based on the few versions of weights determination procedures in the the MWHD formula stated in the current paper. As an application example we employ this formulas for Intuitionistic Fuzzy Distances.

Keywords: Modified weighted Hausdorff distance, Degree of friendship, Intuitionistic fuzzy sets.

Link:

ON THE ALGORITHMIC ASPECT OF THE MODIFIED WEIGHTED HAUSDORFF DISTANCE

Evgeniy Marinov

http://www.foibg.com/ijima/vol03/ijima03-02-p03.pdf

Abstract: Generalized Nets with Characteristics of the Places (GNCP) are conservative extensions of the ordinary Generalized Nets (GNs). In the present paper, for the first time algorithm for transition functioning in GNCP is proposed. Some possible applications of GNCP are discussed. It is shown how GNCP can be used for evaluation of the places and to simplify the graphical structure of the net. In analogy with the concept of reduced GNs, reduced GNCP are introduced. It is proved that there exist two minimal reduced GNCP — one with characteristics of the places and one in which only the tokens obtain characteristics. An example of a reduced GNCP that describes the work of a fire company is presented.

Keywords: Algorithm for transition functioning, Extensions of generalized nets, Generalized nets, Reduced generalized nets.

Link:

REDUCED GENERALIZED NETS WITH CHARACTERISTICS OF THE PLACES

Velin Andonov

http://www.foibg.com/ijima/vol03/ijima03-02-p02.pdf

Abstract: Index Matrices (IMs) are extensions of the ordinary matrices. Some of the operations and relations defined over IMs are analogous of the standard ones, while there have been defined new ones, as well. Operators that do not have analogue in matrix theory have been defined, as well. In general, the elements of an IM can be real or complex numbers, as well as propositions or predicates. In the present paper, IMs with their elements being functions, are defined and some of their properties are discussed.

Keywords: Function, Index matrix, Operation, Operator, Relation.

Link:

INDEX MATRICES WITH FUNCTION-TYPE OF ELEMENTS

Krassimir Atanassov

http://www.foibg.com/ijima/vol03/ijima03-02-p01.pdf

Abstract: The creation the strategy of development of large socio-economic systems is based on the synthesis of the foresight methodology with the methodology of cognitive modeling, which opens up a unique opportunity in the framework of a single program-analytical complex to solve problems of strategic planning and rapid response. The model problem of innovative development of the recreational sector of the Southern coast of the Crimea is considered.

Keywords: foresight methodology, cognitive modeling, sensitivity of the system, structural stability, sustainability of the perturbation, stability by value.

ACM Classification Keywords: H.4.2. INFORMATION SYSTEM APPLICATION: type of system strategy

Link:

THE CREATION OF STRATEGY FOR INNOVATION DEVELOPMENT OF SOCIOECONOMIC SYSTEMS

Vladimir Pankratov

http://www.foibg.com/ijima/vol03/ijima03-01-p08.pdf

Abstract: Scientometric studies on self-citations reveal variety of aspects, like attempted fraud approaching a crime, a self-advertising tool, a standard scientific publication practice that saves space and time by reducing repetitions, and so on up to self-citations being an important element of scientific networking. It seems, however, that self-citation analysis will remain a hot topic for a long time mainly due to its effect on assessment of scientific impact. Therefore, an obvious question that arises is, to what extent self-citations could modify the basic scientometric indicators and indexes, used to perform such assessment. This paper represents an attempt to quantitatively estimate the effect of self-citations on several widely used scientometric indexes (Hirsch’s h, Egghe’s g and Zhang’s e). This has been achieved by incorporating selfcitations into various explicitly given citation-paper rank distributions under more or less realistic assumptions. The latter are deduced making use of well known empirical relationships, based on analyzing a considerable amount of scientometric data. The results obtained contain self-citation corrections for the scientometric indexes, as well as some indications for a ‘normal’ and ‘extraordinary’ self-citation behavior.

Keywords: self-citations, citation-paper rank distributions, scientometric indexes, h-index, g-index, e-index, empirical relationships, scientometric data analysis

ACM Classification Keywords: H. Information Systems, H.2. Database Management, H.2.8. Database applications, subject: Scientific databases; I. Computing methodologies, I.6 Simulation and Modeling, I.6.4. Model Validation and Analysis

Link:

SELF-CITATIONS EFFECT ON SCIENTOMETRIC INDEXES

Vladimir Atanassov, Ekaterina Detcheva

http://www.foibg.com/ijima/vol03/ijima03-01-p07.pdf

Abstract: It is well known that the h-index has been originally introduced by analyzing a simple deterministic model of individual’s scientific activity. The basic features of Hirsch’s model are listed as follows: i) constant publication rate in time; ii) constant number of citations gained by each paper per unit time. One of its predictions is the formation of (approximately) linear negative slope citation-paper rank distribution. However, the actual citation-paper rank distributions obtained by analyzing real life scientometric data are, apart from some rare exceptions, convex and far from linear. In this paper we consider two possible reasons that might explain the deviation from linear negative-slope (Hirsch) citation-paper rank distribution. The first one is an increasing in time publication rate, reflecting the fact of growing productivity in the course of a scientific career, as well as the increase of publication rates on global scale. The second one is an increasing in time number of citations gained by individual papers per unit time, associated, e.g. with improving publication quality or growing popularity. Both factors lead to a shaping of convex citation-paper rank distribution from the linear one, in a way that generalizes Hirsch’s results.

Keywords: citation-paper rank distributions, Hirsch’s model, linear negative-slope distribution, convex distributions, time dependence, empirical relationships, scientometric data analysis

ACM Classification Keywords: H. Information Systems, H.2. Database Management, H.2.8. Database applications, subject: Scientific databases; I. Computing methodologies, I.6 Simulation and Modeling, I.6.4. Model Validation and Analysis

Link:

SHAPING THE CITATION-PAPER RANK DISTRIBUTIONS: BEYOND HIRSCH’S MODEL

Vladimir Atanassov, Ekaterina Detcheva

http://www.foibg.com/ijima/vol03/ijima03-01-p06.pdf

Abstract: In this paper economic order quantity model with two level of trade credit in one replenishment cycle, time value of money, inspection errors, planned backorders, and sales returns is offered. In those problem demand rate, selling price for good quality items, selling price for defective items, interest paid rate and interest earned rate consider as fuzzy variable. Following expected value model criterion and a fuzzy expected value model are constructed. Fuzzy simulations are employed to estimate the expected value of fuzzy variable and maximum return on sale. In order to solve the fuzzy expected value model, a genetic algorithms based on the fuzzy simulation is designed.

Keywords: EOQ, imperfect quality, misclassification errors, deterioration items, inventory backordering, trade credit, fuzzy simulation, genetic algorithm.

ACM Classification Keywords: I.6.5 Model Development

Link:

FUZZY EXPECTED VALUE MODEL WITH INSPECTION ERRORS AND TWO LEVEL OF TRADE CREDIT IN ONE REPLENISHMENT CYCLE

Olha Yegorova

http://www.foibg.com/ijima/vol03/ijima03-01-p05.pdf

Abstract: Original method to study students’ conceptual knowledge was reported at previous ITHEA conferences. According to this method, to show their educational achievements students had to pass through some testing procedure, which requires combining of basic terms from a learnt course into interrelated pairs. Then students’ answers, saved in text files, were analyzed by computer software that joined all interrelated concepts into groups, independent from each other. The larger these groups were the better students’ success in digestion of learnt course was supposed. As human cognitive processes are very complex, it is important to make sure that experimental method is really valid and steady. The present paper examines stability of the experimental results; new implemented experiments were also aimed at estimation of method’s inaccuracy. Several diversified tests were made to assure that method error do not deface observed picture: the difference between repeated tests was evaluated; the stability of student’s pair selection and of knowledge fragmentation were analyzed to make sure that we can trust the results of experiments; regression towards the mean effect, which can skew the outcome of low-performed students, was estimated – the situation was found to be secure; another discipline was checked using the same method and there was no noticeable dependence of numeric scaled gauges from a learning content. Detailed study shows that the generality of students demonstrate a growth of the proposed measure of learning success and its magnitude distinctly exceeds inaccuracy of method. The value of this inaccuracy was also estimated.

Keywords: concept, relation, semantic network, knowledge representation, learning

ACM Classification Keywords: H.1.2 User/Machine Systems – Human information processing, I.2.4 Knowledge Representation Formalisms and Methods, Experimentation – Semantic networks, I.2.6 Learning, Human factors – Concept learning, Knowledge acquisition.

Link:

SEMANTIC NET FROM CONCEPTS AS A MODEL OF STUDENT’S KNOWLEDGE: HOW STABLE ARE THE RESULTS OF EXPERIMENTAL STUDY?

Evgeny A. Eremin

http://www.foibg.com/ijima/vol03/ijima03-01-p04.pdf

Abstract: The article offers information system of time series forecasting based on the combined models of hybrid and selective types for various criteria of selection with the previous time series clustering by the methods of nearest neighbor and K - nearest neighbors.

Keywords: time series, combined forecasting model, information system, K-nearest neighbors’ method.

ACM Classification Keywords: G.3 Probability and statistics - Time series analysis; H.4.2 Information Systems Applications: Types of Systems: Decision Support; H.1.1.Systems and Information Theory

Link:

INFORMATION SYSTEM OF FORECASTING BASED ON COMBINED MODELS WITH TIME SERIES CLUSTERING

Alexander Berzlev

http://www.foibg.com/ijima/vol03/ijima03-01-p03.pdf

Abstract: The article deals with the agent-based approach to creating simulation models. A classification is proposed for agent models depending on their behavior and ability to move. Depending on the ability to move and affect other objects, we have been proposed to distinguish between active and passive agents, spatial and time series agents. Particular attention in this article is paid to time series agents and their simulation using neural networks. The article also considers training and testing tool for neural networks developed within the research for the possibility of using neural networks for time series simulation.

Keywords: agent, agent simulation, rescue neural network, time series.

Link:

SIMULATION OF WEATHER IN AGENT MODELS

Vitaliy Lytvynov, Artem Zadorozhnii

http://www.foibg.com/ijima/vol03/ijima03-01-p02.pdf

Abstract: The inverse problem of choice is formulated as a problem of the settings multiattribute utility model for known-rank alternatives. This problem has the analogy with neural network learning. There is the difference between multiattribute utility model and neural network learning: useful function parameters are subjects for updating instead weight inputs of neurons. It is formulated the main condition alternatives reordering. They must be not comparable in Pareto-dominance analyses. Two alternatives are changed their places due to useful function ratio under its parameters variation. The algorithm is proposed to train the choice model. Algorithm is illustrated by an example.

Keywords: multiattribute utility, inverse problem of choice, utility function, training model, aggregate objective functions.

ACM Classification Keywords: H.4.2 Information Systems Applications: Types of Systems decision support

Link:

NEURAL NETWORK APPROACH TO THE FORMATION MODELS OF MULTIATTRIBUTE UTILITY

Stanislav Mikoni

http://www.foibg.com/ijima/vol03/ijima03-01-p01.pdf

Аннотация: В работе рассмотрен подход к моделированию многокритериальных задач принятия решений с иерархической структурой. Приводится способ согласования целей при помощи модели задачи многокритериального выбора с целевой точкой, координаты которой являются удовлетворяющими значениями оценок по критериям.

Ключевые слова: иерархические структуры, многокритериальный рациональный выбор, точка удовлетворения, функция принадлежности, нечеткое множество.

ACM Classification Keywords: H.4.2 Information Systems Applications: Types of Systems: Decision Support.

Link:

МОДЕЛИРОВАНИЕ СОГЛАСОВАННОСТИ ЦЕЛЕЙ В ИЕРАРХИЧЕСКИХ СТРУКТУРАХ ПРИНЯТИЯ РЕШЕНИЙ

Н.Н. Mаляр

http://www.foibg.com/ijima/vol02/ijima02-04-p10.pdf

Аннотация: Рассматриваются методы поиска решений на основе прецедентов и аналогий в интеллектуальных системах поддержки принятия решений. Исследуется возможность представления прецедентов на основе онтологии предметной области. Предлагается метод поиска решения на основе прецедентов с использованием онтологического подхода и структурной аналогии.

Ключевые слова: интеллектуальные системы поддержки принятия решений, правдоподобные рассуждения, прецеденты, аналогия, онтологический подход.

ACM Classification Keywords: H.4.2 Information systems applications: Types of systems – Decision support; I.2.3 Artificial intelligence: Deduction and Theorem Proving – Uncertainty, "fuzzy," and probabilistic reasoning; I.2.6 Artificial intelligence: Learning – Analogies

Link:

МЕТОД ПОИСКА РЕШЕНИЙ В ИНТЕЛЛЕКТУАЛЬНЫХ СИСТЕМАХ ПОДДЕРЖКИ ПРИНЯТИЯ РЕШЕНИЙ НА ОСНОВЕ ПРЕЦЕДЕНТОВ

Павел Варшавский, Роман Алехин

http://www.foibg.com/ijima/vol02/ijima02-04-p09.pdf