Economy, finances, management: topical issues of science and practical activity

VOLONTYR Liudmyla – Сandidate of Тechnical Sciences, Аssociate Professor, Аssociate Professor of the Department of Мodelling and Іnformation Тechnologies in Economy, Vinnytsia National Agrarian University (21008, Vinnytsia, 3 Soniachna Str., email vm@vsau.vin.ua).

Development of modern economic trends in the system of conceptual foundations for the improvements in sugar beet production sector has necessitated the introduction of new approaches in the processes of managing commodity, financial and information flows on the basis of the use of methods of economic and mathematical modeling. The main idea for implementation these methods is to evaluate the development of forecasts in terms of their formalization, systematization, optimization and adaptation under application of new information technologies.

The quality of management decision-making depends on the accuracy and reliability of the developed long-term evaluations. In this regard, one of the most important areas of research in the economy is to forecast the parameters of the beet industry development and to obtain predictive decisions that form the basis for effective activity in the process of achieving tactical and strategic goals.

Under a significant dispersion of the time series levels, a variety of smoothing procedures are used to detect and distinguish the trend: direct level equalization by the ordinary least squares technique, ordinary and weighted moving averages, exponential smoothing, spectral methods and application of splines, moving average method, or running median smoothing. The most common among them are regular and weighted moving averages and exponential smoothing.

Investigation of methods of forecasting parameters of development of beet growing industry taking into account the peculiarities of constructing quantitative and qualitative forecasts requires solving the following tasks:

- investigation of the specifics of the use of statistical methods of time series analysis in beet growing;

- research of the specificity of the use of forecasting methods for the estimation of long-term solutions in beet growing;

- carrying out practical implementation of the methods as exemplified by the estimation of forecasts of sugar beet yields at the enterprises of Ukraine.

The method of exponential smoothing proposed by R. G. Brown gives the most accurate approximation to the original statistical series – it takes into account the variation of prices. The essence of this method lies in the fact that the statistical series is smoothed out with the help of a weighted moving average, which is subject to the exponential law.

When calculating the exponential value of time t it is always necessary to have the exponential value at the previous moment of time, and therefore the first step is to determine some Sn-1 value that precedes Sn. In practice, there is no single approach to defining initial approximations – they are set in accordance with the conditions of economic research. Quite often, the arithmetic mean of all levels of the statistical series is used as Sn-1.

It should be noted that a certain problem in forecasting with the help of exponential smoothing is the choice of the parameter a optimal value, on which the accuracy of the results of the forecast depends to a large extent. If the parameter a is close to the identity element, then the forecast model takes into account only the effects of the last observations, and if it approaches to zero, then almost all the previous observations are usually taken into account. However, scientific and methodical approaches to determining the optimal value of the smoothing parameter have not yet been developed. In practice, the value of a is chosen according to the smallest dispersion of deviations of the predicted values of the statistical series from its actual levels.

The method of exponential smoothing gives positive findings when a statistical series consists of a large number of observations and it is assumed that the socioeconomic processes in the forecasting period will occur approximately under the same conditions as in the base period.  

A correctly selected model of the growth curve shall correspond to the nature of the trend change of the phenomenon under study. The procedure for developing a forecast using growth curves involves the following steps:

- choice of one or several curves whose shape corresponds to the nature;

- time series changes;

- evaluation of the parameters of the selected curves;

- verification of the adequacy of the selected curves of the process being foreseen;

- evaluation of the accuracy of models and the final choice of the growth curve;

- calculation of point and interval forecasts. 

The most common practice in forecasting are the functions used to describe processes with a monotonous nature of the trend of development and the absence of growth boundaries.

On the basis of the studied models, smoothing of the statistical series of the sugar beet gross yields of in Ukraine was carried out. The statistical data from 1990 to 2017 have been taken for the survey. 

The forecast of the sugar beet yields for 2012-2017 have been used to determine the approximation error by the ordinary moving averages with a length of the smoothing interval of 5 years and 12 years, as well as by the method of exponential smoothing with the parameter α = 0,3 and α = 0, 7

The analysis of the quality of forecasts is based on the average absolute deviation. Therefore, this value is the smallest for the forecast, which is determined by the method of exponential smoothing with the constant value of a = 0,7.

By this method, we will determine the forecast for the next 5 years.

1. Kaletnіk G.M., Olіjnіchuk S.T., Skoruk O.P. (2012). Ekonomіchna efektivnіst' vikoristannya bіoetanolu v Ukrainі [Economic efficiency of bioethanol use in Ukraine]. Zbіrnyk naukovyh prac' VNAU. Serіya: Ekonomіchnі nauki – Collection of scientific papers of VNAU. Series: Economic sciences, no 1(56), pp. 3-7 [in Ukrainian].
2. Fedoseev V.V., Garmash A.N., Dajitbegov D.M. (2001). Ekonomiko-matematicheskie metody i prikladnye modeli [Economic and Mathematical Methods and Applied Models]. Moskow, YUNITI, 391 p. [in Russian].
3. Mur Dzheffri, Uehderford Larri R. (2004). Ekonomicheskoe modelirovanie v Microsoft Excel [Economic Modeling in Microsoft Excel]. Moskow, Izdatel'skij dom «Vil'yams», 1024 p. [in Russian].
4. Granberga A.G. (1990). Stohasticheskoe modelirovanie i prognozirovanie [Stochastic Modeling and Forecasting]. Moskow, Finansy i statistika, 381 p. [in Russian].
5. Fester E.H., Renc B. (1983). Metody korrelyacionnogo i regressionnogo analiza [Methods of correlation and regression analysis]. Moskow, Finansy i statistika, 304 p. [in Russian].
6. Vidy srednih skol'zyashchih [Types of middle moving] (SMA, EMA, WMA). Available at: http://berg.com.ua/tech/indicators-overlays/types-of-moving-averages/ [in Russian].
7. Kremen', V. M., Kremen' O. І. (2014). Fіnansova statistika [Financial statistics]. Kyiv, Centr uchbovoyi lіteratury, 368 p. Avilable at: https://pidruchniki.com/1334020362696/statistika/statistichni_metodi_analizu_fondovogo_rinku [in Ukrainian].
8. Graboveckij B.Ye. (2000). Osnovy ekonomіchnoho prohnozuvannya [Fundamentals of Economic Forecasting]. Vіnnytsia, VF TANG, 209 p. [in Ukrainian].
9. Statistichnij zbіrnik «Sіl's'ke gospodarstvo Ukrainy, 2017 rіk» [Agriculture of Ukraine in 2017]. Available at: www.ukrstat.gov.ua [in Ukrainian].