Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems
In the article represented the new situational-event model of hybrid patterns recognition. This model based on representation a heterogeneous data of a complex system in the form of patterns set, sets of external conditions characteristics as manifestations of a current situation, a static component...
Збережено в:
| Опубліковано в: : | Математичне моделювання в економіці |
|---|---|
| Дата: | 2019 |
| Автор: | |
| Формат: | Стаття |
| Мова: | Англійська |
| Опубліковано: |
Інститут телекомунікацій і глобального інформаційного простору НАН України
2019
|
| Теми: | |
| Онлайн доступ: | https://nasplib.isofts.kiev.ua/handle/123456789/168501 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems / O.I. Zakhozhay // Математичне моделювання в економіці. — 2019. — № 4(17). — С. 16-25. — Бібліогр.: 26 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1859831672523980800 |
|---|---|
| author | Zakhozhay, O.I. |
| author_facet | Zakhozhay, O.I. |
| citation_txt | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems / O.I. Zakhozhay // Математичне моделювання в економіці. — 2019. — № 4(17). — С. 16-25. — Бібліогр.: 26 назв. — англ. |
| collection | DSpace DC |
| container_title | Математичне моделювання в економіці |
| description | In the article represented the new situational-event model of hybrid patterns recognition. This model based on representation a heterogeneous data of a complex system in the form of patterns set, sets of external conditions characteristics as manifestations of a current situation, a static component of a situation - in the form of stationary informativity characteristics, a dynamic component in the form of a nonstationary informativity characteristics and the set of classes as recognitions result. The developed model using provides a priory level of classification reliability, based on analysis of a smaller set but the most informative signs.
В статті представлено нову ситуаційно-подійну модель гібридного розпізнавання образів, засновану на поданні характеристик складної системи у вигляді сукупності образів, множини характеристик зовнішніх умов – як прояву ситуації, статичної складової ситуації – у вигляді множини стаціонарних характеристик інформативності, динамічної складової – у вигляді нестаціонарних характеристик інформативності, та сукупності класів, як апріорно заданих можливих результатів класифікації. Розроблена модель забезпечує отримання апріорно заданого рівня достовірності розпізнавання на основі аналізу меншого набору, але найбільш інформативних даних.
|
| first_indexed | 2025-12-07T15:32:24Z |
| format | Article |
| fulltext |
~ 16 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
UDC 004.93:004.02
https://orcid.org/0000-0002-9078-3242
O. I. ZAKHOZHAY
SITUATIONAL-EVENT MODEL OF THE HYBRID PATTERNS
RECOGNITION FOR HETEROGENEOUS DATA PROCESSING IN
COMPLEX SYSTEMS
Abstract. In the article represented the new situational-event model of
hybrid patterns recognition. This model based on representation a
heterogeneous data of a complex system in the form of patterns set, sets of
external conditions characteristics as manifestations of a current situation,
a static component of a situation – in the form of stationary informativity
characteristics, a dynamic component in the form of a nonstationary
informativity characteristics and the set of classes as recognitions result.
The developed model using provides a priory level of classification
reliability, based on analysis of a smaller set but the most informative
signs.
Keywords: hybrid patterns recognition, making-decision methods, data
classification reliability, time-complexity of recognition algorithms, data
processing in complex systems, program engineering, information systems
and technologies.
DOI: 10.35350/2409-8876-2019-17-4-16-25
Introduction
For data analysis in complex systems, methods and means of pattern recognition
are traditionally used [1-5]. This is due to the impossibility of complete
formalization and mathematical model`s representation of such systems [1, 3, 4, 6].
At the same time, the situation becomes complicated because the large quantity of
data, that characterizing a complex system also are heterogeneity [7, 8]. For data
processing in complex systems the combined recognition is widely used [1, 7, 9-15].
However, the analysis shows, that with heterogeneous data, this approach is not
effective. For combined recognition with a large amount of data, obtaining a
reliable result is associated with a large time complexity [11-14]. That makes
practical application of such algorithms much more difficult. In addition, combined
recognition is not effective with dynamic interference and distortion of the external
environmental condition for the complex system [8, 16, 17].
Hybrid recognition [18], which in some sources is also called multi-parameter
combined recognition [16, 17, 19, 20], are using for data processing in complex
systems, where information patterns have a different nature of origin. It creates a
possibility for getting high-reliability classification result at the changing of
different interferences and distortion in environmental condition for recognition
object. But, for decreasing time-complexity and maximizing accuracy result, the
decision-making should be received by less quantity of information signs. For
complex systems, its aspect is complicated by the heterogeneous data, that to be
processed, also different physical nature of interference and distortion in
environmental condition and dynamical changing their intensity level.
О.І. Захожай, 2019
~ 17 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
Existing approaches to solving that trouble are related with increasing quantity
of informative signs of complex systems. It finds confirmation by the Vapnik-
Chervonenkis deep probabilistic theory of patterns recognition, which is called also
VC-theory [21, 22]. Thus, increasing quality of information signs leads to the level
of reliability is grow. But if the complex system characterized by heterogeneous
data and their informativity level depends on interference and distortion – reliable
classification level will not grow unequivocally with an increase in the number of
signs. The signs affected by interference and distortion will be significant source of
classification errors. Thus, the heterogeneous data of complex systems should be
had maximal informativity level for current interferences and distortion, that have
been present in environmental condition.
According to the aforesaid, creating the new approach to increasing
classification reliability and minimizing time-complexity of heterogeneous data
analysis have a relevance for modern science and technology. The solution to this
problem is seen in the development of a new model of hybrid image recognition.
1. Subject area analysis and statement of the problem
For the heterogeneous data processing in complex systems, is widely used that
model [1, 10]:
{ } { } { } { }d p l sP X X X X= ∪ ∪ ∪ , (1)
where {X}d – set of deterministic, probabilistic, logic and structure (linguistics) in
complex systems signs.
Based on model (1), at the heterogeneous data processing, both direct and
reverse recognition task can have implemented [1, 3, 23]:
{ } { } { } { } { }{ }, d p l sC C C X X X X∀ ∈ ⇔ ∪ ∪ ∪ , (2)
where {C} – classes alphabet, where everyone class C is characterizing the
compliance status of complex system.
The heterogeneous data processing is carried out in the following sequence.
1) The system of the reference set of the algorithm is determined. From signs
set {X}, selecting subsets S1 … Sq had becomes.
2) Calculates the proximity measure of recognition object Or and everyone
representative object with known classification result, and total estimate is
determined:
( ) ( ) ( )
( ) ( ) ( )
1 1 11
1
, , ... , ,
...................................................................
, , ... , ,
r r rS S
r m r m r mS S
q
q
O C O C O C
O C O C O C
Γ = Γ + + Γ
Γ = Γ + + Γ
(3)
where ( ),r mSq O CΓ – compliance estimate for object Or and class Cm by signs
subset Sq.
~ 18 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
3) Next, the rules are specified by which a classification decision can be made
in the following sequence:
– the similarity rules of the reference and analyzed object, which allow
according to a measure estimate to calculate a value that is an estimate for pairs of
objects;
– the rules of estimates forming for each class according to a fixed reference
set based on estimates of pairs for objects;
– the rules of summary estimates forming for each class according to all
reference set;
– the decision-making rules that, based on the estimates for classes, ensure
that the recognition object is classed to one of the classes or, conversely, deny such
classification.
However, for complex systems, the main information is not in individual
characteristics, but in their different combinations [23]. Since in complex systems
it is not always known which combinations are informative, in algorithms of
calculation of estimates the measure of similarity of objects is calculated not by
sequential comparison of individual characteristics, but by comparison of all
possible characteristics that describe a complex system [23].
Thus, combined recognition for analysis of non-uniform data in complex
systems is associated with a large power of reference sets. As results – the large
number of computational operations. In addition, combined recognition does not
take into account the fact that various interferences and distortions result in not all
patterns having the same information value and allow for a reliable solution. In
addition, it should be noted that the effect of interference and distortion, and thus
the information of signs, is dynamic.
Hybrid recognition can take into account the effects of interference and
distortion by describing a complex system with a set of patterns that have different
nature origins [18, 23].
Hybrid recognition uses two approaches: joint patterns analysis [17] and
separate patterns analysis [16].
In joint analysis, all patterns signs of different nature are combined into one
pattern together. After that, the data processing is performed by classical methods
of combined recognition [17]. The positive effect in such a case is to use more
independent numbers of signs on which the interferences and distortions have
different effects.
The maximum benefit of hybrid recognition can be obtained using separate
analysis. In this case, patterns are not merged into a single global pattern and are
processed until the a priori specified recognition reliabilities is obtained to one or
more patterns [16, 18, 23]. This variant is natural for multithreaded data
processing. Given the heterogeneity of the data, different algorithms can be applied
in different stream to match the characteristics and calculate the degree patterns
proximity to representative descriptions of classes for the complex system. The
classification decision-making based on searching of patterns groups with identical
classifications [18]. Since it is a priori known that each image of a complex system
must point to the same class, in the case of an ideal display system and no
interferences and distortion of signs, absolutely reliable result will be obtained. At
the same time, the presence of interference and distortion leads to the facts that part
of the patterns to indicate erroneous classification variants.
~ 19 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
In these circumstances, the possibility of obtaining the most reliable solution
in the minimum time is solely due to the need to exclude from processing less
informative images that have been subject to maximum distortion and
transformation.
Thus, an important scientific and technical task is formed – ensuring a
reliability classification decision-making in hybrid recognition based on analysis of
a smaller set of the most informative signs.
2. Solving problem
For solving task of maximizing recognition reliability and reduction quantity of
less-informativity signs of patterns was developed the new situational-event model
of hybrid patterns recognition which gives possibility for the patterns selecting
with most informativity and create classification result based on matching fewer
data. Thus, the set of data to be processed, varies according to the changes state of
the environmental condition, interferences and distortion level. This model is
represented on cortege such view
, , , ,SEMHPR P ЕС SICh NSICh C= , (4)
where P – a patterns set of complex system, which forms on the multiple
information sources different nature origin;
EC – an environmental condition set that characterized a current situation;
SICh – a patterns stationary informativity characteristics of complex systems;
NSICh – a patterns nonstationary informativity characteristics of complex systems;
C – a classes set, that complex systems characterized.
According to this model, the current state of a complex system is defined on
the basis of receiving set of its images which signs have the different nature of
origin. Besides, receive the external conditions characteristics set allows to define
degree of informational content for each pattern and to provide accept reliable the
decision. For this purpose, on the basis of set EC provide determining of stationary
and non-stationary informativity characteristics for everyone a pattern from {P}.
If to present that pattern of recognition object Р is described by some function
( , )P f x y= , where x, y – arguments, that define characteristic of object, which is
in some space Ω. Then reflection of this object І defined as ( ', ')I g x y= , where x’,
y’ – arguments that define patterns characteristic which is in some information
space Ω’. In the case of an ideal reflective system, for any point of space the
condition satisfied: g f= . In reality, that condition cannot be fully executed. This
is due to the reflection distortion that have a presence. According to relation
between objects space and their patterns space, the reflection of space point will be
defined as:
'( ', ') ( ', ', , , '( , ))g x y h x y f= α β α β , (5)
where (α, β) – point of space`s coordinate
h(x’, y’, x, y) – function, that described the spatial relation between an object and
its reflection.
~ 20 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
Obviously, for other source that have same placement
''( ', ') ( ', ', , , ''( , ))g x y h x y f= α β α β . (6)
On the common situation, ratio of objects space and patterns space for
nonlinear system will have a next form:
,
( ', ') ( ', ', ( , ))
x X y Y
g x y h x y f x y dxdy
∈ ∈
= ∫∫ . (7)
For a linear system:
,
( ', ') ( ', ') ( , )
x X y Y
g x y h x y f x y dxdy
∈ ∈
= ∫∫ . (8)
As at hybrid patterns recognition the signs have the different nature of origin,
value x and y do not correlate with each other and function h will have an
appearance:
( ', ', , ) '( ', ) ( ', )h x y x y h x x h y y= . (9)
This function takes into account the distortion of the object in the pattern
space, and the resulting pattern will be represented as:
,
( ', ', , ) ( , ) ( ', ')
x X y Y
P h x y x y f x y dxdy x y
∈ ∈
= + ξ∫∫ , (10)
where ( ', ')x yξ – the distortions distribution characteristic of objects in pattern
space.
Taking into account (4) and on the basis of (10), at hybrid recognition, the
complex system description model will have an appearance
,
,
( ', ', , ) ( , ) ( ', '),
( ', ', , ) ( , ) ( ', '),
.................................................................................
( ', ', , ) ( , )
1 1 1 1
x X y Y
2 2 2 2
x X y Y
k k k
P h x y x y f x y dxdy x y
P h x y x y f x y dxdy x y
P h x y x y f x y dx
∈ ∈
∈ ∈
= + ξ
= + ξ
=
∫∫
∫∫
,
( ', ').k
x X y Y
dy x y
∈ ∈
+ ξ
∫∫
(11)
Based on model (1), the stationary informativity characteristic for each
patterns of complex system are determined so:
( ', ', , )k
k
1SICh
h x y x y
= , (12)
~ 21 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
where hk – directly, is characteristics of information displaying means (an accuracy
characteristic of sensors for complex systems parameters registration and also
accuracy of data transferring interfaces). As the structure of information system do
not changing in system life-time, these characteristics need only a priori definitions
and are constants in the course of further operation cycles of non-uniform data.
The nonstationary informativity characteristics for each from k patterns are
determined so:
( ', ')k
k
1NSICh
x y
=
ξ
, (13)
where ξ k – is characteristics of the current interferences and distortion for each
patterns of complex system. These characteristics change throughout systems life-
cycle in depending on the current environmental conditions. This characteristics set
is non-stationary and demands constant control on each cycle of classification
decision-making.
The general informativity characteristic of each patterns will be defined as
multiplication of two components: stationary and nonstationary. The most reliable
solution Rd is obtained behind the result of patterns analysis, that satisfies
informational criterion:
{ } , maxk kP P ICh Rd∃ ∈ ⇒ . (14)
The patterns that, on current observation conditions for recognitions object,
have an information characteristic less than a priori caused – mast keep from
recognition. Thus, implementation the situational analysis conception by selected
data that having higher informativity level behind current situations, allowed to
solve task – ensuring the reliability classification decision-making in hybrid
recognition based on analysis of a smaller set of the most informative signs.
The situational-event model of the hybrid patterns recognition, that have been
represented, had find the using on three different applied solution. It was:
information systems of temperature’s spatial distribution monitoring for coke pie
[24], information system of ultrasonic linear distance measurement for automation
systems [25] and information system of text-unique level analysis [26]. In these
solutions, mean an 18% reduction in the amount of data processed was obtained for
a priori given level of recognition reliability of 85%. This effect is clarified by the
use of situational processing of the most informative data under current
environmental conditions. Thus, the main problem was solved.
Conclusion
The main results presented in it work consist in the following.
1. For efficient application of recognition, it is necessary to provide
processing of less but more informative data in the current situation and level of
interference and distortion.
2. For ensuring a priory defined reliability of recognition at simultaneous
decrease a time-complexity of the heterogeneous data analysis, the new situational-
~ 22 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
event model of hybrid patterns recognition was developed. That model considers
the level of stationary and non-stationary informativity characteristics of complex
systems patterns was developed. In this case, decision-making happens on the basis
of processing smaller quantity, however the most informative data (at the current
level of hindrances and distortions as manifestations of a situation).
The efficiency of the offered model was confirmed experimentally for three
various information systems: analysis of spatial distribution of temperature of coke
pie, ultrasonic measurement of linear distances and verifications of text data on
uniqueness. For all three applications, a positive result was obtained that confirmed
the correctness of the problem solution.
REFERENCES
1. Simankov V. S., Lutsenko E. V., Adaptive management of complex systems based on
the theory of pattern recognition. Krasnodar: Techn. univ. Kuban state technol. univ., 1999. –
318 p. (in Russian)
2. Leigh W., Purvis R., Ragusa J.M., Forecasting the NYSE composite index with
technical analysis, pattern recognizer, neural network, and genetic algorithm: a case study
in romantic decision support. Decision Support Systems. – 2002. – vol. 32, – pp. 361-377.
(in English)
3. Yang-Yu Liu, Albert-Laszlo Barabasi, Control principles of complex systems. Review of
modern physics. – 2016. – vol. 88. – iss. 3., – 58 p. DOI: 10.1103/RevModPhys.88.035006.
(in English)
4. Kyriakos G. Vamvoudakis, Sarangapani Jagannathan, Control of complex systems.
Elsevier Inc., 2016. – 762 p. DOI:10.1016/C2015-0-02422-4 (in English)
5. Wen Q., Yang Z., Song Y., Jia P., Automatic stock decision support system based on
box theory and SVM algorithm. Expert systems with Applications. – 2010. – vol. 37. –
pp. 1015-1022. (in English)
6. Fang Le, Yang Fan, Dong Wen, Guan Tong, Expectation propagation with stochastic
kinetic model in complex interaction systems. 31st International Conference on Neural
Information Processing Systems. – 2017. – pp. 2026-2036. (in English)
7. Menyaylenko О. S., Zakhozhay O. I., Synthesis fundamentals of classifiers for technical
systems of patterns recognition with the use of human’s models of emotional processes.
Scientific Bulletin of National Mining University. – 2015. – vol. 1 (145), – pp. 120-126. (in
Ukrainian)
8. Yang Fan, Liu Bo, Dong Wen, Optimal Control of Complex Systems through Variational
Inference with a Discrete Event Decision Process. 18th International Conference on
Autonomous Agents and Multiagent Systems. – 2019. – pp. 296-304. (in English)
9. Zhuravlev Yu. I, Ryazanov V. V., Senko O. V., Recognition. Mathematical methods.
Program System. Practical using. Moscow: Phazis, 2005 – 159 p. (in Russian)
10. Niemann H., Zhuravlev Yu., Gourevitch I., Laptev I., Pattern recognition and image
understanding. Amsterdam: IOS Press, 2000. – 340 p. (in English)
11. Barabash Yu. L., Collective statistic solving in time recognition. Moscow: Radio and
communication, 1983. – 224 p. (in Russian)
12. Gorelik A. L., Skripkin V. A., Recognition methods. Ed. 2. Moscow: High School,
1984. – 219 p. (in Russian)
13. Garcia-Pedrajas N., Ortiz-Boyer D., Improving multicast pattern recognition by the
combination of two strategies. IEEE Transactions on Pattern Analysis and Machine
Intelligence. – 2006. – vol. 28. – pp. 1001-1006. (in English)
14. Eduard C. Popovici, Liviu A. Stancu, Ovidiu G. Guta, Stefan C. Arseni, Octavian Fratu,
Combined use of pattern recognition algorithms for keystroke-based continues
authentication system. 10th International Conference on Communications (COMM). –
2014. – pp. 91-109. (in English) DOI: 10.1109/ICComm.2014.6866686
~ 23 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
15. Tsypkin Ya. Z., Base of information technology of identification. Moscow: Science,
1984. – 520 p. (in Russian)
16. Zakhozhay O. I., Menyaylenko A. S., Lyfar V. A., Separate Analysis of Informational
Signs in Multi-Parametric Combined Patterns Recognition Systems. Problemele Energeticii
Regionale. – 2019. – vol. 1-1 (40), – pp. 60-68. (in Russian) DOI: 10.5281/zenodo.3239140
17. Zakhozhay O. I., Joint analysis of informational signs in multi-parametric combined
patterns recognition systems. Electrotechnical and computer systems. – 2019. – vol. 29
(105). – pp. 78-86. (in Ukrainian)
18. Zakhozhay O. I., Lyfar V. O., Ivanov V. G., The data consolidation and decision
making by the ranking of groups with identical classification in hybrid recognition systems.
Electrotechnical and computer systems. – 2019. – vol. 31 (107). – pp. 95–103. (in
Ukrainian)
19. Zakhozhay О. І., Criteria for the determination of informativity and patterns ranking for
making decisions in multi-parametric combined recognition systems. Electrotechnical and
computer systems. – 2015. – vol. 27 (103). – pp. 196-204. (in Ukrainian)
20. Zakhozhay O. I., Lyfar V. O., Baturin O. I., The decision making by the finding of
groups with identical classification in multiparametric combined patterns recognition
systems. Bulletin of Volodymyr Dahl East Ukrainian University. – 2019. – vol. 9 (249). –
pp. 36-42. (in Ukrainian)
21. Cherkassky V., Mulier F., Vapnik-Chervonenkis (VC) learning theory and its
applications. IEEE Transactions on Neural Networks. – 1999. – vol. 10. – iss. 5. – pp. 985-
987. (in English) DOI: 10.1109/TNN.1999.788639
22. Devroye L., Györfi L., Lugosi G., Vapnik-Chervonenkis Theory. A Probabilistic
Theory of Pattern Recognition. Stochastic Modelling and Applied Probability. – 1996. –
vol. 31. – pp. 187-213. (in English) DOI: 10.1007/978-1-4612-0711-5_12
23. Zakhozhay O. I., The concept of improving the theoretical and methodological
foundations of synthesis information technologies for automated information processing
and complex systems controlling. Bulletin of Volodymyr Dahl East Ukrainian University. –
2018. – vol. 6 (246). – pp. 49-55. (in Ukrainian)
24. Menyaylenko O. S., Zakhozhay O. I., The combined systems of patterns recognition in
monitoring of temperature’s spatial distribution for coke pie. Electrotechnical and computer
systems. – 2017. – vol. 12 (88). – pp. 147-154. (in Ukrainian)
25. Menyaylenko A. S., Zakhozhay O. I., The information technology of automated data
processing in the multi-channel ultrasonic measurement systems. Research Bulletin of the
National Technical University of Ukraine "Kyiv Polytechnic Institute". – 2014. – vol. 6
(98). – pp. 62-67. (in Ukrainian)
26. Menyaylenko O. S., Zakhozhay O. I., Bidyuk P. I., The text-unique verification
reliability increasing by using of combined patterns recognition systems. System research
and information technologies. – 2017. – vol. 4. – pp. 29-37.
СПИСОК ЛІТЕРАТУРИ
1. Симанков В. С. Адаптивное управление сложными системами на основе теории
распознавания образов / В. С. Симанков, Е. В. Луценко. – Краснодар: Техн. унив.
Кубан. гос. технол. унив., 1999. – 318 p.
2. Leigh W. Forecasting the NYSE composite index with technical analysis, pattern
recognizer, neural network, and genetic algorithm: a case study in romantic decision
support / W. Leigh, R. Purvis, J. M. Ragusa // Decision Support Systems. – 2002. – vol. 32, –
pp. 361-377.
3. Yang-Yu Liu. Control principles of complex systems / Liu Yang-Yu, Barabasi Albert-
Laszlo // Review of modern physics. – 2016. – vol. 88. – iss. 3., – 58 p. DOI:
10.1103/RevModPhys.88.035006.
~ 24 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
4. Kyriakos G. Vamvoudakis. Control of complex systems / G. Vamvoudakis Kyriakos,
Jagannathan Sarangapani // Elsevier Inc., 2016. – 762 p. DOI:10.1016/C2015-0-02422-4
5. Wen Q. Automatic stock decision support system based on box theory and SVM
algorithm / Q. Wen, Z. Yang, Y. Song, P. Jia // Expert systems with Applications. – 2010. –
vol. 37. – pp. 1015-1022.
6. Fang Le. Expectation propagation with stochastic kinetic model in complex interaction
systems / Le Fang, Fan Yang, Wen Dong, Tong Guan // 31st International Conference on
Neural Information Processing Systems. – 2017. – pp. 2026-2036.
7. Меняйленко О. С. Основи синтезу класифікаторів технічних систем розпізнавання
образів з використанням моделей емоційних процесів людини / О. С. Меняйленко,
О. І. Захожай // Науковий вісник Національного гірничого університету. – 2015. –
№ 1 (145). – С. 120-126.
8. Yang Fan. Optimal Control of Complex Systems through Variational Inference with a
Discrete Event Decision Process / Fan Yang, Bo Liu, Wen Dong // 18th International
Conference on Autonomous Agents and Multiagent Systems. – 2019. – pp. 296-304.
9. Журавлев Ю.И. Математические методы. Программная система. Практические
применения / Ю.И. Журавлев, В. В. Рязанов, О. М. Сенько. – Москва: Фазис, 2005. –
159 с.
10. Niemann H. Pattern recognition and image understanding. / H. Niemann,
Yu. Zhuravlev, I. Gourevitch, I. Laptev. Amsterdam: IOS Press, 2000. – 340 p.
11. Барабаш Ю. Л. Коллективные статистические решения при распознавании /
Ю. Л. Барабаш. – Москва: Радио и связь. – 1983. – 224 с.
12. Горелик А. Л., Скрипкин В. А. Методы распознавания. Изд. 2 / А. Л. Горелик,
В. А. Скрипкин. – Москва: Высшая школа, 1984. – 219 p.
13. Garcia-Pedrajas N. Improving multicast pattern recognition by the combination of two
strategies / N. Garcia-Pedrajas, D. Ortiz-Boyer // IEEE Transactions on Pattern Analysis
and Machine Intelligence. – 2006. – vol. 28. – pp. 1001-1006.
14. Eduard C. Popovici. Combined use of pattern recognition algorithms for keystroke-
based continues authentication system / C. Popovici Eduard, A. Stancu Liviu, G. Guta
Ovidiu, C. Arseni Stefan, Fratu Octavian // 10th International Conference on
Communications (COMM). – 2014. – pp. 91-109. DOI: 10.1109/ICComm.2014.6866686
15. Ципкин, Я. З. Основы информационной технологии идентификации /
Я. З. Ципкин. – Москва: Наука, 1984. – 520 p.
16. Захожай О. И. Раздельный анализ информационных признаков в
многопараметрических комбинированных системах распознавания образов /
О. И. Захожай, А. С. Меняйленко, В. A. Лыфарь // Проблемы региональной
энергетики. – 2019. – № 1-1 (40). – С. 60-68. DOI: 10.5281/zenodo.3239140
17. Захожай О. І. Спільний аналіз інформаційних ознак в багатопараметричних
комбінованих системах розпізнавання образів / О. І. Захожай // Електротехнічні та
комп’ютерні системи. – 2019. – № 29 (105). – С. 78-86.
18. Захожай О. І. Консолідація даних і прийняття рішень на основі ранжування груп
ідентичних класифікацій в гібридних системах розпізнавання / О. І. Захожай,
В. О. Лифар, В. Г. Іванов // Електротехнічні та комп’ютерні системи. – 2019. –
№ 31 (107). – С. 95-103.
19. Захожай О. І. Критерії визначення інформативності та ранжування образів при
прийнятті рішень в багатопараметричних комбінованих системах розпізнавання /
О. І. Захожай // Електротехнічні та комп’ютерні системи. – 2015. – № 27 (103). –
С. 196-204.
20. Захожай О. І. Прийняття рішень на основі пошуку груп ідентичних класифікацій в
багатопараметричних комбінованих системах розпізнавання образів / О. І. Захожай,
В. О. Лифар, О. І. Батурін // Вісник Східноукраїнського національного університету
імені Володимира Даля. – 2019. – № 9 (250). – С. 36-42.
~ 25 ~
Математичне моделювання в економіці, №4, 2019. ISSN 2409-8876
21. Cherkassky V. Vapnik-Chervonenkis (VC) learning theory and its applications /
V. Cherkassky, F. Mulier // IEEE Transactions on Neural Networks. – 1999. – vol. 10. –
iss. 5. – pp. 985-987. DOI: 10.1109/TNN.1999.788639
22. Devroye L. Vapnik-Chervonenkis Theory. A Probabilistic Theory of Pattern
Recognition / L. Devroye, L. Györfi, G. Lugosi // Stochastic Modelling and Applied
Probability. – 1996. – vol. 31. – pp. 187-213. DOI: 10.1007/978-1-4612-0711-5_12
23. Захожай О. І. Концепція вдосконалення теоретико-методологічних основ синтезу
інформаційних технологій автоматизованої обробки інформації і управління
складними системами / О. І. Захожай // Вісник Східноукраїнського національного
університету імені Володимира Даля. – 2018. – № 6 (246). – С. 49-55.
24. Меняйленко О. С. Комбіновані системи розпізнавання образів при аналізі
просторового розподілу температури коксового пирога / О. С. Меняйленко,
О. І. Захожай // Електротехнічні та комп’ютерні системи. – 2013. – № 12 (88). –
С. 147-154.
25. Меняйленко О. С. Інформаційна технологія автоматизованої обробки даних в
багатоканальних системах ультразвукового вимірювання / О. С. Меняйленко,
О. І. Захожай // Наукові вісті Національного технічного університету України
“Київський політехнічний інститут”. – 2014. – № 6 (98). – pp. 62-67.
26. Меняйленко О. С. Підвищення достовірності перевірки унікальності текстів
шляхом використання комбінованих систем розпізнавання образів /
О. С. Меняйленко, О. І. Захожай, П. І. Бідюк // Системні дослідження та інформаційні
технології. – 2017. – № 4. – С. 29-37.
Received on 01.10.2019.
10. Niemann H., Zhuravlev Yu., Gourevitch I., Laptev I., Pattern recognition and image understanding. Amsterdam: IOS Press, 2000. – 340 p. (in English)
10. Niemann H. Pattern recognition and image understanding. / H. Niemann, Yu. Zhuravlev, I. Gourevitch, I. Laptev. Amsterdam: IOS Press, 2000. – 340 p.
|
| id | nasplib_isofts_kiev_ua-123456789-168501 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 2409-8876 |
| language | English |
| last_indexed | 2025-12-07T15:32:24Z |
| publishDate | 2019 |
| publisher | Інститут телекомунікацій і глобального інформаційного простору НАН України |
| record_format | dspace |
| spelling | Zakhozhay, O.I. 2020-05-04T11:52:15Z 2020-05-04T11:52:15Z 2019 Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems / O.I. Zakhozhay // Математичне моделювання в економіці. — 2019. — № 4(17). — С. 16-25. — Бібліогр.: 26 назв. — англ. 2409-8876 DOI: 10.35350/2409-8876-2019-17-4-16-25 https://nasplib.isofts.kiev.ua/handle/123456789/168501 004.93:004.02 In the article represented the new situational-event model of hybrid patterns recognition. This model based on representation a heterogeneous data of a complex system in the form of patterns set, sets of external conditions characteristics as manifestations of a current situation, a static component of a situation - in the form of stationary informativity characteristics, a dynamic component in the form of a nonstationary informativity characteristics and the set of classes as recognitions result. The developed model using provides a priory level of classification reliability, based on analysis of a smaller set but the most informative signs. В статті представлено нову ситуаційно-подійну модель гібридного розпізнавання образів, засновану на поданні характеристик складної системи у вигляді сукупності образів, множини характеристик зовнішніх умов – як прояву ситуації, статичної складової ситуації – у вигляді множини стаціонарних характеристик інформативності, динамічної складової – у вигляді нестаціонарних характеристик інформативності, та сукупності класів, як апріорно заданих можливих результатів класифікації. Розроблена модель забезпечує отримання апріорно заданого рівня достовірності розпізнавання на основі аналізу меншого набору, але найбільш інформативних даних. en Інститут телекомунікацій і глобального інформаційного простору НАН України Математичне моделювання в економіці Інформаційні технології в економіці Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems Ситуаційно-подійна модель гібридного розпізнавання образів для обробки неоднорідних даних в складних система Article published earlier |
| spellingShingle | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems Zakhozhay, O.I. Інформаційні технології в економіці |
| title | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems |
| title_alt | Ситуаційно-подійна модель гібридного розпізнавання образів для обробки неоднорідних даних в складних система |
| title_full | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems |
| title_fullStr | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems |
| title_full_unstemmed | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems |
| title_short | Situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems |
| title_sort | situational-event model of the hybrid patterns recognition for heterogeneous data processing in complex systems |
| topic | Інформаційні технології в економіці |
| topic_facet | Інформаційні технології в економіці |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/168501 |
| work_keys_str_mv | AT zakhozhayoi situationaleventmodelofthehybridpatternsrecognitionforheterogeneousdataprocessingincomplexsystems AT zakhozhayoi situacíinopodíinamodelʹgíbridnogorozpíznavannâobrazívdlâobrobkineodnorídnihdanihvskladnihsistema |