CC BY
39
23. Neural crest and mesoderm lineage-dependent gene expression in orofacial development / Bhattacherjee V., Mukhopadhyay P., Singh S., Johnson C., Philipose J. T., Warner C. P. et. al. // Differentiation. 2007. Vol. 75, Issue 5. P. 463-477. doi: 10.1111/j.1432-0436.2006.00145.x
24. Harrington J. The desirability function // Industrial Quality Control. 1965. Vol. 21, Issue 10. P. 494-498.
25. ARACNE: An Algorithm for the Reconstruction of Gene Regulatory Networks in a Mammalian Cellular Context / Margolin A. A., Nemenman I., Basso K., Wiggins C., Stolovitzky G., Favera R., Califano A. // BMC Bioinformatics. 2006. Vol. 7, Issue (Suppl 1). P. S7. doi: 10.1186/1471-2105-7-s1-s7
26. Fawcett T. ROC graphs: Notes and practical considerations for researches // Machine learning. 2004. Vol. 31, Issue 1. P. 1-38.
27. Kaiser S. Biclustering: Methods, software and application. Munchen, 2011. 178 p.
Пропонуються адаптивн комбтован ModeMÏ гiбридного та селективного munie для прогнозування часових рядiв на основi про-грамного набору з адаптивних nолiномiаль-них моделей рiзнuх порядтв. Пропонуються адаптивн комбтован моделi прогнозування часових рядiв з врахуванням результа-miв iденmuфiкацiï nодiбносmей в ретроспек-ци цих часових рядiв. Оцтена ефективтсть прогнозування рiзнuх комбтованих моделей залежно вiд рiвня nерсuсmенmносmi часових рядiв. Розроблен моделi дозволяють тдви-щити точтсть у випадку середньостроко-вого прогнозування нестащонарних часових рядiв, зокрема фтансових показнитв
Ключовi слова: прогнозування часових рядiв, пошук nодiбносmей, адаптивна ^m6ï-
нована модель, показник Герста
□-□
Предлагаются адаптивные комбинированные модели гибридного и селективного типов для прогнозирования временных рядов на основе программного набора из адаптивных полиномиальных моделей разных порядков. Предлагаются адаптивные комбинированные модели прогнозирования временных рядов с учетом результатов идентификации подобий в ретроспекции этих временных рядов. Оценена эффективность прогнозирования различных комбинированных моделей в зависимости от уровня персистентности временных рядов. Разработанные модели позволяют повысить точность в случае среднесрочного прогнозирования нестационарных временных рядов, в частности финансовых показателей
Ключевые слова: прогнозирование временных рядов, поиск подобий, адаптивная комбинированная модель, показатель Херста
UDC 004:519.2
|dOI: 10.15587/1729-4061.2018.121620|
DEVELOPMENT OF ADAPTIVE COMBINED MODELS FOR PREDICTING TIME SERIES BASED ON SIMILARITY IDENTIFICATION
A. Kuchansky
PhD, Associate Professor Department of Cybersecurity and Computer Engineering*
E-mail: kuczanski@gmail.com A. Biloshchytskyi
Doctor of Technical Sciences, Professor, Head of
Department
Department of Information Systems and Technologies Taras Shevchenko National University of Kyiv Volodymyrska str., 60, Kyiv, Ukraine, 01033 Yu. Andrashko Lecturer
Department of System Analysis and Optimization Theory
Uzhhorod National University Narodna sq., 3, Uzhhorod, Ukraine, 88000 S. Biloshchytska PhD, Associate Professor Department of Information Technology Designing and
Applied Mathematics* Ye. Shabala PhD, Associate Professor Department of Cybersecurity and Computer Engineering*
O. Myronov Postgraduate student Department of Information Technologies* *Kyiv National University of Construction and Architecture Povitroflotskyi ave., 31, Kyiv, Ukraine, 03037
1. Introduction
It is known that the overwhelming majority of financial, technical and physical processes for which the problem of predicting arises are characterized by nonlinearity and
instability with respect to the average level. Application of classical econometric prediction models and appropriate methods of predicting such time series that reflect these processes is rather limited. This is because of a low efficiency of these models in such conditions. Prospective directions
©
in development of information systems for analysis and prediction of time series and related decision-making support systems include construction of combined multi-level models and application of methods for intelligent data analysis. In addition, the pre-prediction evaluation of input data including financial indicators that can be implemented on the basis of fractal R/S analysis is relevant.
Contradictions between the requirements to planning strategies of activities of financial institutions and the tasks of scientifically grounded decision-making support can be solved by the use of mathematical models for prediction of financial time series. Known mathematical models do not ensure a necessary precision in predicting time series for solving problems of financial risk management under uncertainty. In this connection, the studies that consist in solving this contradiction by development of combined adaptive models using the results of intelligent analysis of time series, in particular, identification of similarities by the methods of the nearest neighbor and application of the apparatus of fractal analysis of data for pre-predictive analysis of time series should be considered relevant.
2. Literature review and problem statement
The problem of identifying similarities in the dynamics of time series is an important component in the study of changes in stock exchange price values, search for the goods with similar characteristics in terms and conditions of implementation. Besides, the problem of identifying similarities is used to analyze similarity of acoustic, textual, numeric or graphical information, detection of anomalies or events in signals of different nature, etc.
In [1], the concept of "approximately similar time series data" was developed which ensures understanding similarity between time series and is the basis for identification of similarities. However, this concept does not include a method for calculating similarity between time series by routine calculation of metric distances between them. The classical foundations of the problem of identifying similarities and their classification have been described in [2, 3]. In [4], basic theses of the problem of search for similarities are described by comparison with a sample, the methods of calculating metric distances, the problem of segmentation and approximation of time series are described.
Numerous papers [5-26] describing application of similarity identification in various applied problems can be mentioned. In particular, predictive power of the clustering method using similar behavior of the stock market time series as well as the use of this method for efficient prediction of share prices is investigated in [5]. The method for modeling samples for the problem of short-term prediction of time series is proposed in [6]. Work [7] considers the method of selective comparison with the sample. This method is used for the problem of constructing combined models of prediction of increment signs of time series with an unstable nature of fluctuations taking into account identification of similarities or indexation. The method for prediction of increments of time series in conditions of uncertainty using trend models of fluid averages is described in [8]. In papers [9, 10], the method of the closest neighbor is considered for this problem but not the comparison with the sample. The method of indexing time series based on identification of similarities between them is described in [11]. This method
is effective for time series of different lengths. The method of clustering for the search for similarities in multidimensional spaces represented by multidimensional vectors or time series of a fixed length is described in [12].
The use of similarity identification methodologies is also useful in education to predict potentials of development of scientific directions. The method that ensures calculation of prediction of the potential of scientific directions is described in [13]. The values of time series in this case are the ratio of estimations of the results of scientific research activities of scientists for various periods of time. The method can be used to identify prospective directions of research that are formed in the scientific medium. The method for classifying scholars by the research directions based on identification of similarities is described in [14]. The method for constructing evaluations of the results of scientific research activities of scientists based on the analysis of citations of publications is described in [15]. Paper [16] describes a parametric model for assessing and predicting quality of educational institutions which uses approaches to comparison of grades. The method enabling assessment of not only educational institutions but also structural units of higher educational institutions using transformation of a complex of developed estimations from qualitative to quantitative ones is described in paper [17]. The task of identifying similarities in this case may be useful for constructing relevant assessments taking into account other educational institutions or structural units of a similar weight.
The similarity identification task is also used to detect incomplete duplicates in text information for the purpose of preventing plagiarism and illegal dissemination of information. In particular, authors of work [18] have developed a method for detecting incomplete duplicates in tables which is based on the methods of the nearest neighbor and locally sensitive hashing. In work [19], a conceptual model of the system for finding incomplete duplicates using identification of similarities in electronic documents is described. The model enables identification of incomplete duplicates in documents containing data of various types: text information, numeric data, tables, mathematical formulas, schemes, diagrams and other graphic images.
For predicting time series, in addition to identification of similarities, conventional models can also be used. However, it should be borne in mind that traditional econometric models often show results with a significant margin of error, especially when applied to financial time series. This is explained by the fact that in the conditions of crisis phenomena, financial time series are often weakly persistent or even close to random, that is, they lose the memory of their initial conditions. Nevertheless, if incoming time series are trend-resistant, the use of traditional prediction models makes sense. In this case, the main stages in the prediction construction are [7]: predictive retrospection, construction of a model for prediction and formalization of the model, construction of the prospectus, estimation of the prediction and its verification.
Paper [20] describes approaches to intellectual decision-making methods in business using classical models of time series prediction. Adaptive models of short-term prediction of time series and methods of constructing combined methods are described in [21]. Models and methods for prediction of time series using intelligent data analysis: neural network modeling, genetic algorithms, fuzzy analysis, etc. are considered in work [22]. All these methods can use the
.«¡T
concept of identifying similarities and, as a result, increase their effectiveness.
Both classical prediction models and the prediction models using identification of similarities are often used as components of complex prediction, modeling and decision-making systems. Analysis of the peculiarities of designing information-analytical systems for prediction of time series with the use of expert evaluation has been carried out in work [23]. In particular, the method of fuzzy clustering which uses the identification problem of similarities is considered in [24]. The use of expert estimation to predict time series in information systems is also described in [25]. Peculiarities of project configuration management in development of distributed systems which can use identification of similarities between projects according to the set of indicators that characterize these projects are considered in [26].
A promising attempt is combination of traditional prediction models and the method of identifying similarities in retrospection of time series and constructing combined prediction models. These models should be endowed with adaptive properties to dynamically adapt to the mechanisms that generate time series. It is assumed that such models will be effective in predicting time series in conditions of uncertainty, in particular, data from stock markets, commodity time series, etc.
3. The aim and objectives of the study
The study objective was to create combined models of prediction of time series in conditions of uncertainty with adaptive characteristics at a maximum efficiency. Under the maximum efficiency of a model, it is understood its achieving the minimum average prediction errors.
To achieve this objective, the following tasks were set:
- construct adaptive combined models of hybrid and selective types for prediction of time series;
- construct adaptive combined models of hybrid and selective types for prediction of time series taking into account the results of identification of similarities in the retrospection of these time series.
4. Formal statement of the problem of prediction of time series and estimation of prediction results
Let {zi }"0 is some discrete time series of length n with no spaces:
{z, L ={zo, V", zn } =
= {z(to),z(tl),.-,z(tn)}, (1)
where ^ eS are the discrete moments of time in which the values of time series are fixed, i = 0, n, S is a discrete set, t0 is the initial time moment [7].
Based on the retrospective values zn, zn+i,..., zn.m+1 of the time series {zi }"o, n>m, estimate most accurately behavior of this time series in the future at time moments tn+1, tn+2,..., tn+0, i. e. construct a sequence of predictive values:
{Z }n=n+1 ={ Zn+1,Zn+2,"', Zn+e}, (2)
where 0 is the prediction horizon and m is the volume of the retrospective sample.
Let zt (n) be the prediction which is calculated at the moment tn (at the point n), t points ahead, t = 1, e. Functional dependence which allows one to describe behavior of the time series is called the prediction model. Denote by F the model that describes behavior of the time series {zi} . Prediction of this time series for one point ahead (t=1) based on the model F can be written as:
Zn+1 = 21 (n) = F (Zn-m+1, Zn—m, • ■ ■, Zn ), (3)
where z1 (n) is the prediction of the time series calculated at the point n, one point ahead.
In the case of prediction with the horizon t>1, we can write:
Zn+1 = Z1 (n) = F (Zn-m+1, Zn—m+2, • ■ ■, Zn ), Zn+2 = Z2 (n) = F (Zn-m+2, Zn—m+3, • ■ ■, Zn, Zn+1),
zn+e = ze(n) = F (zn-m+e,•■■, Zn+e-2, Zn+e—1), (4)
where m is the volume of the retrospective sample, n>m.
As a rule, prediction models have a number of parameters to be evaluated before calculating the prediction. These parameters can be of adaptive nature, i. e., they change dynamically taking into account the prediction errors in the previous steps. The model that satisfies the respective criterion of prediction quality estimation will be considered the most accurate. Such criteria can be a mean square error, a mean absolute deviation, a standard deviation, a relative error, etc.
Most models require a certain amount of retrospective information to calculate the prediction. If sufficient historical data are available for time series observation, it is expedient to estimate quality of the prediction model in a given time series before calculating the prediction. Estimates can be used to construct confidence intervals of prediction or make more accurate parameters of the prediction model.
To evaluate quality of the prediction model F, consider a retrospective time series Z' = {zi }= of length m+1. Construct predictive values of the retrospective time series based
on a series {Zn—m—h+1, Zn-m-h+2,•■■, Zn-m-1} °f h+1 length, denoting
it as: Z" = {z }n m 1 , ,. To evaluate prediction quality, 1, 2,..., 0
L i J i=n—m—h+1
points ahead, iteratively apply the model F according to the scheme (4).
For t=1, this can be written formally as follows:
21 (n — m — 1) = F(Zn-m-h+1, Zn-m-h+2,; Zn-m-1),
(n — m) = F ( Zn—m+h+2, Zn—m+h+3,-', Zn—m ), Z1 (n — 1) = F (Zn—m—1,•, Zn—2, Zn—1 ^ m + l £ n (5)
where z1 (i — 1) is the prediction made at the moment i-1 for one point ahead, that is, prediction of the i-th element of the retrospective series Z', i = n — m, n. The sequence of such
predictions is denoted by Z[ = {21 (i — 1)}n .
For t=2, the process of finding prediction estimates will have the form:
21 (n — m — 1) = F ( Zn—m—h+1, Zn—m—h+2,• ", Zn—m—1),
z2 (n — m — 1) = F(Zn—m+h+2, Zn—m+h+3, •, Z1 (n — m — 1)),
Z4 (n m ) — F ( -n—m—h+2, Zn-m-h+3 Zn-m ), Z2 (n " m) — F (Zn-m+h+3 , Zn-m+h+f' ? (n — m)),
(n — 2) — F ( Zn-m-2- Zn-m-V'- Zn-2 )-^ (n — 2) — F(zn—m—1- Zn—m w ? (n — 2))- m + / ^ (6)
where ?2 (i — 2) is the prediction made at the moment of i-2, 2 points ahead, that is, the prediction of the i-th element of the retrospective series Z', i — n — m +1- n. The sequence of
such elements is denoted by Z'2 — {z2 (i — 2)}" Using the
above scheme, other sequences of predictions Z3', Z', etc. can be constructed.
To construct a criterion of assessing quality of the prediction model, it is necessary to determine which of the predicted values of each of the sequences Z', t — 1,0 should be taken into account in the target function of the criterion. Call the subsequences of the sequences Z', t — 1,0, used to estimate accuracy of prediction evaluative and denote by
Z'*—{^(i — t)L, ,
where Jt are the sets of indexes of elements of sequences Z', chosen to construct the evaluative sequence Z'* t — 1,0, n-m+T-1<i<n, card(JT)<m, card(JT) is the number of elements in sequences Z'*. The sequence Z'*, t — 1,0 can be constructed of the most significant predictions using expert judgment. Besides, a quantitative estimate of significance of each individual prediction can be indicated with the help of weighting factors. That is, all prediction values Z'*— Z', t — 1,0, will be included in the resulting target function, but with the weighting factors that can take zero values.
After construction of the evaluative series, criterion of quality of the prediction model is calculated which enables determining of the degree of satisfaction of the objectives that were set. In this case, the objective is to achieve maximum accuracy of predictions.
Consider the main criteria of quality assessment:
(11) respectively which are calculated according to the
formulas t —1,0, ^ra; — 1:
i—1
2
E — E (^Z'*) — -L ^ (V — -') , (8)
' i=1,V'
JJ
E2=E2 Z ¿T*) = fI», ( 2T - z; )\ (9)
E3 — E3 (z',Z'*) — — YraJ--^x 100%, (10)
jeJ,
E' — E' (Z', Z'*) — max ra — (11)
jeJ,
can be used for estimation.
Let us estimate quality of L prediction models of F type that were tested in a retrospective series. Denote by Z' J the evaluative series obtained by the i-th model in predicting a retrospective series Z', t steps ahead at p — 1, L. Then, the model that corresponds to the minimum deviation of the prediction values from the real
Ej (Z',Z;p*min, j — 0,4, p — %L
will be considered to be optimal one for each model. One of the approaches that can take into account the benefits of the models from a program set at different sections of the time series, is construction of combined models for prediction of selective and hybrid types.
5. Combined adaptive prediction models taking into
account similarities in the retrospection of time series
E0 = E 0 IZ', Z f) = —I
(7)
- prediction on the basis of the mean absolute deviation taking into account significance of predictions using the normalized weights o,, i — 1, vT,
Ira, — 1,
i—1
here and thereafter, n-m+t-1<j<n, vT=card(JT), vT<m, t — 1,0. The notation
ET° = E0 (Z Z T
T \ ' T
Let the program set 3PS of the prediction models f f2,..., fL be given on the basis of which estimates of future elements of the time series can be constructed for the series {- }n at the point n
{2 P
2p 2p }
(12)
where êp+T is the prediction calculated at the point n, t points ahead, according to the p-th model.
It is necessary to build the most accurate sequence of prediction values based on the set 3PS and retrospective values of the time series {zi}
Z = {z. }"+" ={ 2
L 1J i=n+1 I
e}.
means that the criterion E0, is used for estimation according to which the members of the retrospective series Z are compared with the corresponding members of the estimation series Z'*, with the prediction calculated, t steps ahead t — 1,0. Also, mean square error (8), standard deviation (9), average relative error (10) and maximum absolute error
Let a set of non-reference histories of the same length m is constructed for the time series {zi}._0. Denote this set through X. Non-reference (N, m)-histories are such sections of the time series {zi}._0 of fixed length which are constructed by the method of the flow window according to the rule:
) ={zN,Z
-1v, ZN
-l} = {
N + ; } ;
T i=1
N = 0, n — m,
(13)
where the upper index in z^ defines the point from which construction of the history begins and the lower index in the brackets is the length of the history or the length of the section of the time series. In other words, the history of the time series refers to its subsequence of retrospective values or retrospection of a fixed length. Details on the construction of non-reference and reference histories are described in [7].
Denote predictions calculated on the basis of each model fp from the program set %PS, p = 1, L at points m, m+1,..., n-1, m<n of series {zi} , corresponding to the last elements of non-reference histories z^ (13), N = 0, n - m, are the lengths m t = 1,8 points ahead by:
zf(m + N -1) = fp ({zn+j C)
% ( m + N -1) = fp ({zn+j j, zf ( m + N -1)),
zep ( m + N -1) =
Denote it with 3BS. This set is formed at each point of the time series in the process of prediction and consists of such models that give the most accurate predictions at the current section of the 3BS c 3PS. time series. Selection of models for the main set can be done, for example, on the basis of the D-criterion [21]:
Dp (t) <XDmin(t),
(16)
where c is the prehistory period, x is the criterion parameter, Dmln(t) = min Dp (t),
p=1,L
I c 2
DP(t) = -K^P(m-t-j)-zm-j) .
Calculate the value of the B-criterion for each model for the time moment n by the formula:
BPt = (1—^b ) BV
'( j — t) — z
(17)
= f ({*
ZN+j j',zf(m + N — 1),2f(m + N — 1),...,zepi(m + N — 1)), (14)
where zf (m + N -1) is the predictive value of the point zm+N-1+t, following after (N, m)-history obtained on the basis
of the p-th prediction model. Record fp ({zN+¡}. o) means
that the corresponding non-reference (N, m)-history belonging to the set of non-reference histories zNm) was used to construct the prediction according to the model fp.
Construct the reference (N, m)-history for the input time
senes {z, L:
Z(m) {Zn—m+1, Zn— m+2,••■, Zn } {zj }
(15)
and basing on some degree of proximity, for example, using the Euclid distance by the method of the nearest neighbor, find from the set of non-reference histories one that is most similar to the reference one.
An non-reference history z^ e^ is called the most similar to the reference history z^^1, if there are no other non-reference histories z^, k e[1, n - m -1], k x, for which
d(Z'T1,Zk ,)<d(Z"—m+1
-I (m
, Z(
where Bv- is the value of the B-criterion calculated
it -
at the moment j = n - c -—n, c is the prehistory period for each model fp, p = 1, L from the set 3PS and for each value of the prediction period t = 1,8, aB is the smoothing parameter. The problem of determining optimal parameter aB in the B-criterion does not differ from the estimation of the smoothing parameter of the exponential model [20].
Compare the resulting finite values Bpt with each other.
For each t = 1,0, find the minimum value of the B-criterion
for each of the models of the program set. Denote by
min Bpt
P=U
fx the models to which minimum values of B-criteria for each t = 1,0, fv e3PS correspond. Then, the prediction at the point n according to the model ACSM-B is calculated by the formula:
Zt(n) = zt(n),
(18)
where d is some metric distance.
In accordance with the principles of constructing combined models, there are two approaches to calculation of predictions: selective and hybrid. Let us first consider the usual combined models of selective and hybrid types and then the same models but using similarity identifications.
Adaptive Combined Selective Model by B-criterion of selection (ACSM-B)
Let 3PS be a program set of prediction models. The selective approach consists in a selection for each value of t from the program set of models 3PS, a single model which provides high accuracy of prediction according to a certain selection criterion: B-criterion, K-criterion [21], etc. As a rule, parameters of selection criteria are adaptive. Besides, to improve accuracy of prediction, selection criteria are often applied not to the program but to the so-called main set.
where z* (n) are the predictions according to the models that are calculated at the point n, t = 1,8 steps ahead.
Adaptive Combined Selective Model according to the R-criterion of selection (ACSM-R)
Introduce coefficient xp for each model from the program set. Let zn be the last point of the time series Z at which selection of models is performed. Consider the interval [n-c, n] where c is the period of prehistory. At each point in this section, selection of the most accurate models is performed according to the D-criterion (16). If the model fp satisfies the D-criterion, then the coefficient xp of the model fp-xp increments by one. The model with coefficient Xp at the time n being maximal is chosen to calculate the prediction according to the adaptive composite model. If there are several models with the same maximum coefficients, then one of them is selected that gets the coefficient increment later than others. After coming of the new zn+1 point and transition to the next interval [n+1-c, n+1], all coefficients xp, p = 1,L are zeroed and calculation starts from the beginning.
Let the prediction period t=1 and the moment for which values of the D-criterion are calculated is tn.p. The sequence of coefficients xp for each model fp from the program set p = 1, L is given. Assume that before the algorithm is execut-
ed, Xp=0. The general formula for calculating the weighted D-criterion at an arbitrary time moment t for the prediction period t is DP (t) is as follows:
Dp (t) =
\-i
I I(c - (t-t-j)-z
(19)
where c is the period of prehistory.
The first step of the algorithm is calculation of the value Dp(t) for each model p = 1,L. Selection from the models fp those for which the following condition is fulfilled.
Dp (t)£(l.2 + (t-1)-A)-Dm ,„ (t),
where Dmin (t) = minDp (t), is the threshold
h =
1.9 -1.2 c -1 .
According to the ACSM-P model, prediction is calculated like for the previous models by formula (18) where 2* (n) is the prediction, t steps ahead, at the point n according to the model for which the value np is maximal. Adaptive Combined Hybrid Model (ACHM) Let formula (16) be used to select the most exact models at the point n. That is, for each t, the main set 3BS was constructed. Denote by f1t,f,...,ft the prediction models
LB
included in the set 3BS,
ILB is the number of models for each
(20)
fixed t, L%b < L.
Let the values Bpt be calculated at the point n by formula (17) for each of the models fp, p = 1,LXB from the set 3BS and for each value of the prediction 3BS period. Prediction by the model of ACHM will be determined by the formula:
p=1
(n),
(24)
The coefficients of the models fp for which condition (20) is fulfilled increment by one: X = Xp + 1, Xp is the coefficient of the model fp at time t. If r^0, accept r=r-1 and proceed to the first step of the algorithm, that is, continue to calculate the D-criterion for the moment t=n-r-1. If r=0, then the algorithm is complete and the prediction value for t=can be calculated. Denote by f ^ the model to which the maximum value corresponds among the coefficients Xnp, for the given t, f Te3PS, and by 2* (n) the prediction according to this model which is calculated at the point n. Like for the previous model, the prediction is determined by formula (18).
If several models with the same maximum coefficients are selected, then the values of the coefficients for the points preceding the point n: n-1, n-2... are compared. This is necessary to find the model with its coefficient incremented at the time closest to n.
Next, zero coefficients xp=0, increase the value of t, t=2 and proceed to the first step. Continue to execute the algorithm until predictions for each given t = 1,8 are calculated at time n.
It should be noted that the parameter r in this model should be small (r=3) but can be adjusted before implementation of prediction. Use of this selection criterion can increase prediction accuracy compared to other approaches.
Adaptive Combined Selective Model according to the P-criterion of selection (ACSM-P)
Consider the P-criterion of selection. Let zn be the last point of the time series. Like for the R-criterion, consider the section [n-c, n] and calculate for Vj = n - c, n the value
where the weighted coefficients wp are determined on the basis of values Bpt, and the following condition is fulfilled
Lb
I w t = 1.
i=1
Algorithm for choosing weight coefficients in the ACHM model:
1. Form a matrix G = (gj) , consisting of numbers of all models of prediction of a program set 3PS, L is the number of models. For example, if L = 6, matrix G has the form:
2 3 4 5 6
1 3 4 5 6
G = 1 2 4 5 6
1 2 3 5 6
1 2 3 4 6
2. Form the basic sets of models at the point n for each value t. Re-enumerate all models of these sets from 1 to LXB, where L%B = card(3BS). Calculation of weight coefficients is performed according to the formulas:
LB 1 ' 1
=n II Bg„: j=1 i=1
j = 1, IB, t = 1, e,
f —
IS p=1,L,
I fj
j=1
(25)
(26)
Sp = sign ( zp ( j )-Zj
where 2f ( j) is the prediction calculated at the point j, one point ahead, according to the model fp, p = 1, L. Then, according to this criterion at time n, the model taking the maximum value of np is selected:
^ p-77—^ max,
lp p=1,I
where
np =n
V1
v Z y
(21) where gij are the elements of the G matrix.
Adaptive Combined Selective Model with Similarity Identification (ACSMwSI)
Suppose that on the basis of some degree of proximity, the history z^-"*1 eK, x e[m, n -0-1], most similar to the reference history z"""*1 was determined by the method of the nearest neighbor. The last element of history zj"-"*1 will
(22) be the element zx. Construct the main sets of prediction models 3BS, at this point using the D-criterion. That is, models f1t,f,..., ft, LB <L are obtained. For details of choice of similar histories see the work [7].
For each model of the basic set 3BS, calculate the value of
(23) the B-criterion by the formula:
b%=(1 -a,) (x -T),
where 0<aB<1 is the smoothing parameter,
(x-t) = ziT (x-Tl-Z
(27)
(28)
is the absolute error of the prediction calculated at the moment tx.T, t steps ahead, according to the models
and the old reference history z"m™+1 becomes non-reference, that is, it is included in the set X, and the calculation process starts from the beginning. That is, based on a certain degree of proximity, the history most similar to the reference history is found, sets 3BS are formed for each t, the value of the B-cri-terion is calculated for each prediction model that uses the most similar non-reference history as retrospective information. Next, a combined prediction is constructed.
f, t = 1, 6, q = 1, LB.
Denote by fT* the model selected from the main set according to some selection criterion for a fixed value t (27), (28). The prediction calculated on the basis of models fT* at the point n, t steps ahead, denote by Z* (n).
If a model fT* with a minimum value of the B-criterion
min {BJT} corresponds to each t, then the prediction ac-
q=1,ZB
cording to the combined model of selective type according to the B-criterion of selection (27) using the identification of similarities by the method of the nearest neighbor will be calculated by the formula:
z„ (n) = a z„ (n) +
(1 -a)
(29)
where zx+T is the value of the time series which follows after the history z(m~)m+\ most similar to the reference history z^p1, a e [0,1] is a certain parameter.
Adaptive Combined Hybrid Model with Similarity Identification (ACHMwSI)
Apply a hybrid approach to construct a prediction. The prediction according to the hybrid approach is calculated as a weighted sum of predictions for all models that make up the main set 3BS. Let following the identification of similarities at the zx point, basic sets 3BS were formed for each t, and the value of the B-criteria Bq\, q = 1,U„ was calculated.
x,T' It ' B
Denote by Zq (n) the prediction calculated at the point n, t points ahead, according to the models fq from the main set 3BS, q = 1, ZZB, t = 1, 0. Then, the prediction according to the hybrid model with identification of similarities by the method of the nearest neighbor is determined by the formula:
) = a| <
?T=1
-(1 -a
(30)
I® q= 1,
?,=1
that is, by the formulas.
In simple words, prediction in the described combined models is calculated in accordance with the appropriate hybrid or selective principle taking into account the retrospective values of the time series which followed after the corresponding, most similar history.
In the case of dynamic prediction, a new reference history is constructed after calculation of the prediction at point zn and the arrival of a new point zn+1.
z?T2 = {z
' —1 L n-
(m
2 , Zn-m+3 Z П-1 , ZU , Z П+1 } ,
6. Discussion of the results obtained in the study of predicting time series based on the adaptive combined models
For the experiment, time series of raw material prices (daily data) were taken for the period from 2014 to 2017 (by 900 points). The developed models were implemented and tested on these time series. Tables 1-4 show the results of testing some of the selected time series: prices for zinc, copper, platinum, nickel, silver, aluminum. These time series were loaded and pre-processed using sequential R/S analysis to determine the Hurst index. More information on the procedure of fractal sequential R/S analysis can be found in [27]. Moreover, the theoretical index which corresponds to the truth of the basic hypothesis of randomness of these time series,
E
R±
S
m
= 0.540
and was calculated by formula [28]:
E
'R л
S
m
m -1
л/лГ
(31)
where ae[0,1], weights can be determined on the basis of the B-criterion taking into account the coefficient of proportionality determined from the equality of the weights to one
where m is the length of the time series, Rm is the swing, and Sm is the mean root square deviation of the input time series Z of length m. The results of prediction are shown in the Tables 1-6 (the minimum average relative prediction errors for each т value are shown in bold with underlining). The average relative errors were calculated by formula (10). Designation of models in Tables 1-6: SESM-0, -1, -2 are Brown adaptive polynomial models of the 0.1 and 2 orders (Simple exponential smoothing model), ACSM-B, -R, -P are adaptive combined models of selective type with B, R, P selection criteria, ACHM is adaptive combined hybrid model, ACSMwSI-B, -R is adaptive combined selective model with similarity identification in the retrospection of the time series and selection according to B and R criteria, ACHMwSI is adaptive hybrid model with identification of similarities in the retrospection of the time series. The program set of combined models was formed on the basis of SESM models of 0, 1 and 2 orders with various parameters of smoothing.
It can be concluded that in the case of short-term prediction of time series, the maximum accuracy is obtained by the adaptive hybrid model with formation of the basic set 3BS according to the D-criterion with adaptation of the value X and the combined selective models. In the case of prediction for a medium period т>3, an increase in the accuracy of combined models with identification of similarities is observed. In addition, the more persistent the input time series, the more precise the ACHM model. For less persistent
series, prediction accuracy of ACSMwSI-B, ACSMwSI-R and ACHMwSI models is greatest already for t>3. For example, the Hurst index for time series of aluminum prices is H=0.818, that is, this time series is persistent according to the results of a sequential R/S analysis. For other time series, whose prediction results are shown in Tables 1-5, the Hurst index varies in the range of He[0.570, 0.670] which indicates
a weak persistence of these time series. The lowest Hurst index is observed for the time series of prices for zinc H=0.570 (Table 5). Also, as a result of the study, it was found that in the case of sufficiently persistent time series, at H>0.75 it would be sufficient to use conventional adaptive combined models without identification of similarities in the retrospection of these time series.
Table 1
Average relative error of prediction of prices for copper with the horizon t = 1,10
T Models 1 2 3 4 5 6 7 8 9 10
SESM-0 1.683 2.568 3.246 3.785 4.224 4.545 4.826 5.056 5.201 5.258
SESM-1 1.479 2.355 3.068 3.726 4.253 4.654 4.880 5.156 5.469 5.614
SESM-2 1.667 2.353 2.998 3.599 4.135 4.585 4.889 5.186 5.551 5.736
ACSM-B 1.667 2.579 3.330 3.703 4.173 4.615 4.729 5.008 5.390 5.526
ACSM-R 1.670 2.482 3.114 3.566 4.132 4.604 4.856 5.136 5.446 5.631
ACSM-P 1.655 2.349 2.977 3.540 4.075 4.525 4.827 5.124 5.538 5.754
ACHM 1.351 2.270 2.966 3.603 4.101 4.501 4.776 5.045 5.356 5.510
ACSMwSI-B 1.696 2.594 3.057 3.740 4.013 4.167 4.502 4.761 4.941 5.163
ACSMwSI-R 1.916 2.571 3.129 3.560 3.951 4.231 4.548 4.847 5.008 4.996
ACHMwSI 1.491 2.426 2.993 3.502 3.822 4.168 4.448 4.751 4.981 5.079
Table 2
Average relative error of prediction of prices for platinum with the horizon t = 1,10
T Models 1 2 3 4 5 6 7 8 9 10
SESM-0 1.203 1.783 2.251 2.457 2.715 3.061 3.353 3.570 3.701 3.806
SESM-1 1.190 1.887 2.349 2.545 2.703 3.005 3.400 3.665 3.838 3.935
SESM-2 1.434 2.165 2.578 2.819 2.913 3.066 3.473 3.790 3.975 4.044
ACSM-B 1.370 2.007 2.431 2.650 2.811 3.204 3.567 3.691 3.709 3.764
ACSM-R 1.373 1.870 2.398 2.578 2.920 3.132 3.467 3.674 3.849 3.887
ACSM-P 1.363 2.021 2.426 2.661 2.842 3.103 3.475 3.773 4.006 4.096
ACHM 1.082 1.776 2.245 2.506 2.651 2.987 3.360 3.596 3.766 3.842
ACSMwSI-B 1.840 2.271 2.418 2.378 2.559 2.842 2.992 3.099 3.178 3.312
ACSMwSI-R 1.918 2.248 2.605 2.659 2.605 2.812 3.129 3.175 3.178 3.270
ACHMwSI 1.470 2.117 2.345 2.431 2.494 2.780 3.022 3.186 3.227 3.306
Table 3
Average relative error of prediction of prices for nickel with the horizon t = 1,10
T Models 1 2 3 4 5 6 7 8 9 10
SESM-0 2.507 3.711 4.366 4.827 5.077 5.360 5.540 5.643 5.813 6.002
SESM-1 2.353 3.707 4.550 5.034 5.411 5.737 5.944 6.160 6.259 6.425
SESM-2 2.692 4.098 4.974 5.503 5.833 6.016 6.194 6.398 6.523 6.752
ACSM-B 2.832 3.837 4.415 5.098 5.642 5.866 5.787 5.887 6.046 6.250
ACSM-R 2.621 3.604 4.525 4.968 5.575 5.639 5.742 6.087 6.232 6.567
ACSM-P 2.627 3.970 4.727 5.334 5.822 6.016 6.134 6.331 6.319 6.570
ACHM 2.167 3.559 4.332 4.890 5.301 5.599 5.733 5.934 5.995 6.188
ACSMwSI-B 3.025 3.730 4.273 4.773 5.102 5.320 5.514 5.620 5.751 5.919
ACSMwSI-R 3.003 3.859 4.569 4.776 4.847 5.076 5.193 5.213 5.911 6.106
ACHMwSI 2.337 3.769 4.343 4.693 4.980 5.277 5.388 5.551 5.727 5.901
Table 4
Average relative error of prediction of prices for silver with the horizon t = 1,10
^^^^ T Models 1 2 3 4 5 6 7 8 9 10
SESM-0 1.705 2.566 3.228 3.747 4.119 4.422 4.682 5.038 5.345 5.615
SESM-1 1.444 2.434 3.144 3.618 4.035 4.349 4.658 4.904 5.153 5.485
SESM-2 1.726 2.589 3.240 3.679 3.993 4.313 4.670 4.986 5.212 5.395
ACSM-B 1.723 2.666 3.093 3.651 4.033 4.276 4.608 5.065 5.271 5.287
ACSM-R 1.620 2.580 3.101 3.578 3.998 4.297 4.683 5.091 5.344 5.399
ACSM-P 1.626 2.500 3.160 3.635 3.922 4.262 4.650 4.952 5.170 5.420
ACHM 1.311 2.348 3.079 3.566 3.968 4.284 4.605 4.893 5.211 5.456
ACSMwSI-B 1.593 2.506 3.153 3.562 3.967 4.299 4.569 4.860 5.109 5.202
ACSMwSI-R 1.612 2.457 3.047 3.549 3.892 4.285 4.569 4.860 5.109 5.323
ACHMwSI 1.233 2.345 3.005 3.438 3.827 4.167 4.457 4.705 4.979 5.170
Table 5
Average relative error of prediction of prices for zinc with the horizon t = 1,10
^^^^ T Models ^^^^ 1 2 3 4 5 6 7 8 9 10
SESM-0 2.507 3.711 4.366 4.827 5.077 5.360 5.540 5.643 5.813 6.002
SESM-1 2.353 3.707 4.550 5.034 5.411 5.737 5.944 6.160 6.259 6.425
SESM-2 2.692 4.098 4.974 5.503 5.833 6.016 6.194 6.398 6.523 6.752
ACSM-B 2.832 3.837 4.415 5.098 5.642 5.866 5.787 5.887 6.046 6.250
ACSM-R 2.621 3.604 4.525 4.968 5.575 5.639 5.742 6.087 6.232 6.567
ACSM-P 2.627 3.970 4.727 5.334 5.822 6.016 6.134 6.331 6.319 6.570
ACHM 2.167 3.559 4.332 4.890 5.301 5.599 5.733 5.934 5.995 6.188
ACSMwSI-B 3.025 3.730 4.273 4.773 5.102 5.320 5.514 5.320 5.751 5.919
ACSMwSI-R 3.003 3.859 4.869 5.262 4.847 5.076 5.193 5.213 5.911 6.106
ACHMwSI 2.337 3.769 4.343 4.693 4.980 5.277 5.388 5.551 5.727 5.901
Table 6
Average relative error of prediction of prices for aluminum with the horizon t = 1,10
T Models 1 2 3 4 5 6 7 8 9 10
SESM-0 1.538 2.254 2.779 3.190 3.527 3.855 4.266 4.604 4.903 5.224
SESM-1 1.357 2.076 2.630 3.127 3.471 3.807 4.134 4.533 4.884 5.209
SESM-2 1.578 2.217 2.700 3.175 3.531 3.810 4.161 4.496 4.876 5.159
ACSM-B 1.512 2.028 2.556 3.077 3.616 4.060 4.450 4.675 4.824 5.065
ACSM-R 1.512 2.028 2.547 3.220 3.513 4.031 4.287 4.537 4.830 5.179
ACSM-P 1.541 2.238 2.704 3.200 3.546 3.794 4.144 4.469 4.852 5.143
ACHM 1.293 2.017 2.575 3.094 3.441 3.775 4.120 4.476 4.815 5.094
ACSMwSI-B 2.213 2.650 3.148 3.429 3.748 4.063 4.333 4.700 5.071 5.167
ACSMwSI-R 2.198 2.708 3.144 3.400 3.684 4.036 4.334 4.682 5.009 5.144
ACHMwSI 1.757 2.554 3.050 3.349 3.644 4.005 4.291 4.618 4.950 5.197
The described adaptive combined models with similarity identification can be used to predict both stationary and non-stationary equidistant time series without spaces, t steps ahead. It was found that greater accuracy of these models was observed for t>3, that is, in the case of medium-term prediction. For these models, the basic requirement is a sufficient amount of retrospective information, that is, the input series must have at least 700 points.
7. Conclusions
For solution of the problem of prediction of time series, formalization was made and new adaptive combined predic-
tion models of non-stationary time series were proposed, as well as methods for estimating accuracy of prediction. The results of the study are as follows:
1. Adaptive combined selective models of prediction according to B-, R-, P-selection criteria with automatic formation of the basic set of models based on the adaptive D-criterion (ACSM-B, ACSM-R, ACSM-P models) were constructed. The adaptive combined hybrid model of prediction of time series (ACHM model) was constructed.
2. Adaptive combined selective models of prediction according to the R- and B-criteria of selection with identification of similarities in the retrospection of time series by the method of the nearest neighbor (ACSMwSI-R, ACSM-wSI-B models) were constructed. The adaptive combined
hybrid model of prediction with identification of similarities in the retrospection of time series (ACHMwSI model) was constructed.
Adaptive characteristics of combined models with similarity identification in retrospection of time series, compared with conventional combined models, can improve accuracy in the case of medium-term prediction of non-stationary time series, in particular financial indicators. As a result of the experiment, it was found that in the case of short-term prediction for t<2, the ACHM model has the highest
accuracy. ACSM models with various selection criteria are effective in predicting persistent time series with the Hurst index H>0.75 for t>2. In the case of prediction of time series with the Hurst index
Em S
V m y
< H < 0.75
for t>2, the ACSMwSI models with various selection criteria and ACNmwSI are more precise.
References
1. Goldin D. Q., Kanellakis P. C. On similarity queries for time-series data: Constraint specification and implementation // Lecture Notes in Computer Science. 1995. P. 137-153. doi: 10.1007/3-540-60299-2_9
2. Agrawal R., Faloutsos C., Swami A. Efficient similarity search in sequence databases // Lecture Notes in Computer Science. 1993. P. 69-84. doi: 10.1007/3-540-57301-1_5
3. Fast similarity search in the presence of noise, scaling, and translation in time-series databases / Agrawal R., Lin K.-I., Sawhney H. S., Shim K. // VLDB. 1995. P. 490-501.
4. Similarity Measures and Dimensionality Reduction Techniques for Time Series Data Mining / Cassisi C., Montalto P., Aliotta M., Cannata A., Pulvirenti A. // Advances in Data Mining Knowledge Discovery and Applications. 2012. doi: 10.5772/49941
5. Nayak, R., te Braak P. Temporal pattern matching for the prediction of stock prices // 2nd International Workshop on Integrating Artificial Intelligence and Data Mining (AIDM 2007). 2007. P. 95-103.
6. Singh S. Pattern modelling in time-series forecasting // Cybernetics and Systems. 2000. Vol. 31, Issue 1. P. 49-65. doi: 10.1080/019697200124919
7. Kuchansky A., Biloshchytskyi A. Selective pattern matching method for time-series forecasting // Eastern-European Journal of Enterprise Technologies. 2015. Vol. 6, Issue 4 (78). P. 13-18. doi: 10.15587/1729-4061.2015.54812
8. Berzlev A. A method of increment signs forecasting of time series // Eastern-European Journal of Enterprise Technologies. 2013. Vol. 2, Issue 4 (62). P. 8-11. URL: http://journals.uran.ua/eejet/article/view/12362/10250
9. Perlin M. S. Nearest neighbor method // Revista Eletrônica de Administraçao. 2007. Vol. 13, Issue 2.
10. Fernández Rodríguez F., Sosvilla Rivero S. J., Andrada Félix J. Nearest-Neighbour Predictions in Foreign Exchange Markets // SSRN Electronic Journal. 2002. doi: 10.2139/ssrn.300404
11. Kahveci T., Singh A. Variable length queries for time series data // Proceedings 17th International Conference on Data Engineering. 2001. doi: 10.1109/icde.2001.914838
12. Li C., Chang E., Garcia-Molina H., Wiederhold G. Clustering for approximate similarity search in high-dimensional spaces // IEEE Transactions on Knowledge and Data Engineering. 2002. Vol. 14, Issue 4. P. 792-808. doi: 10.1109/tkde.2002.1019214
13. Biloshchytskyi A., Kuchansky A., Andrashko Y., Biloshchytska S., Dubnytska A., Vatskel V. The method of the scientific directions potential forecasting in infocommunication systems of an assessment of the research activity results // 2017 4th International Scientific-Practical Conference Problems of Infocommunications. Science and Technology (PIC S&T). 2017. doi: 10.1109/info-commst.2017.8246352
14. Biloshchytskyi A., Kuchansky A., Andrashko Y., Biloshchytska S., Kuzka O., Shabala Y., Lyashchenko T. A method for the identification of scientists' research areas based on a cluster analysis of scientific publications // Eastern-European Journal of Enterprise Technologies. 2017. Vol. 5, Issue 2 (89). P. 4-11. doi: 10.15587/1729-4061.2017.112323
15. Biloshchytskyi A., Kuchansky A., Andrashko Y., Biloshchytska S., Kuzka O., Terentyev O. Evaluation methods of the results of scientific research activity of scientists based on the analysis of publication citations // Eastern-European Journal of Enterprise Technologies. 2017. Vol. 3, Issue 2 (87). P. 4-10. doi: 10.15587/1729-4061.2017.103651
16. Otradskaya T., Gogunskii V., Antoshchuk S., Kolesnikov O. Development of parametric model of prediction and evaluation of the quality level of educational institutions // Eastern-European Journal of Enterprise Technologies. 2016. Vol. 5, Issue 3 (83). P. 12-21. doi: 10.15587/1729-4061.2016.80790
17. Biloshchytskyi A., Myronov O., Reznik R., Kuchansky A., Andrashko Yu., Paliy S., Biloshchytska S. A method to evaluate the scientific activity quality of HEIs based on a scientometric subjects presentation model // Eastern-European Journal of Enterprise Technologies. 2017. Vol. 6, Issue 2 (90). P. 16-22. doi: 10.15587/1729-4061.2017.118377
18. Lizunov P., Biloshchytskyi A., Kuchansky A., Biloshchytska S., Chala L. Detection of near dublicates in tables based on the locality-sensitive hashing method and the nearest neighbor method // Eastern-European Journal of Enterprise Technologies. 2016. Vol. 6, Issue 4 (84). P. 4-10. doi: 10.15587/1729-4061.2016.86243
19. Biloshchytskyi A., Kuchansky A., Biloshchytska S., Dubnytska A. Conceptual model of automatic system of near duplicates detection in electronic documents // 2017 14th International Conference The Experience of Designing and Application of CAD Systems in Microelectronics (CADSM). 2017. doi: 10.1109/cadsm.2017.7916155
20. Vercellis C. Business intelligence: data mining and optimization for decision making. Cornwall: John Wiley & Sons, 2009. 417 p. doi: 10.1002/9780470753866
21. Lukashin Yu. P. Adaptive methods of near-term time series forecasting: manual. Moscow: Finance and Statistics, 2003. 416 p.
22. Snytyuk V. E. Forecasting. Models. Methods. Algorithms: manual. Kyiv: Maklaut, 2008. 364 p.
23. Mulesa O., Geche F., Batyuk A., Buchok V. Development of combined information technology for time series prediction // Advances in Intelligent Systems and Computing. 2017. Vol. 689. P. 361-373. doi: 10.1007/978-3-319-70581-1_26
24. Mulesa O., Geche F., Batyuk A. Information technology for determining structure of social group based on fuzzy c-means // 2015 Xth International Scientific and Technical Conference "Computer Sciences and Information Technologies" (CSIT). 2015. doi: 10.1109/stc-csit.2015.7325431
25. Mulesa O., Geche F. Designing fuzzy expert methods of numeric evaluation of an object for the problems of forecasting // Eastern-European Journal of Enterprise Technologies. 2016. Vol. 3, Issue 4 (81). P. 37-43. doi: 10.15587/1729-4061.2016.70515
26. Morozov V., Kalnichenko O., Liubyma I. Managing projects configuration in development distributed information systems // 2017 2nd International Conference on Advanced Information and Communication Technologies (AICT). 2017. doi: 10.1109/aiact.2017.8020088
27. Peters E. E. Fractal market analysis: applying chaos theory to investment and economics. John Wiley & Sons Inc., 1994. 336 p.
28. Anis A., Lloyd E. The expected value of the adjusted rescaled Hurst Range of independent normal summands. Biometrika. 1976. Vol. 63, Issue 1. P. 111-116. doi: 10.2307/2335090
-□ □-
Робота видноситься до областi вери-фжацп тбернетичних оцточних показ-нитв, зокрема до виршення завдання визначення обмеженого класу операцш з розподтеними параметрами, об'ек-ти якого мають наперед визначену рейтингову ефективтсть. В рамках дослидження визначаються правила формування операцшних об'eктiв, як можуть використовуватися для тес-тування оцточних показнитв на предмет можливостi гх використання в якостi формули ефективностi
Ключовi слова: верифшащя оцточ-ного показника, операщя з розподшени-ми параметрами, клас операцш, метод
верифшаци
□-□
Работа относится к области верификации кибернетических оценочных показателей, в частности к решению задачи определения ограниченного класса операций с распределенными параметрами, объекты которого имеют предопределенную рейтинговую эффективность. В рамках исследования определяются правила формирования операционных объектов, которые могут использоваться для тестирования оценочных показателей на предмет возможности их использования в качестве формулы эффективности
Ключевые слова: верификация оценочного показателя, операция с распределенными параметрами, класс операций, метод верификации -□ □-
UDC 007.5
[DOI: 10.15587/1729-4061.2018.121810|
DEVELOPMENT OF TEST OPERATIONS WITH DIFFERENT DURATION IN ORDER TO IMPROVE VERIFICATION QUALITY OF EFFECTIVENESS FORMULA
I. L u t se n ko
Doctor of Technical Sciences, Professor* E-mail: delo-do@i.ua E. Fomovskaya PhD, Associate Professor, Head of Department** E-mail: 2fill.fo@gmail.com E. Vi h rova PhD, Associate Professor Department of Mathematics and techniques of its teaching Kryvyi Rih State Pedagogical University Gagarina ave., 54, Kryvyi Rih, Ukraine, 50086 E-mail: elvihrova65@gmail.com O. Serdi uk Assistant** E-mail: olgajs28@gmail.com F. Fomovsky PhD, Associate Professor** E-mail: vladislav.mospan@gmail.com *Department of Information and Control Systems*** **Department of Electronic Devices*** ***Kremenchuk Mykhailo Ostrohradskyi National University Pershotravneva str., 20, Kremenchuk, Ukraine, 39600
1. Introduction
The importance of solving the problems of identification of operations in terms of their effectiveness is related to the
need to maximize the pace of an enterprise development and investment capabilities of its owners [1].
In turn, the use of the effectiveness formula as an optimization criterion [2] makes it possible to formalize and
©
