CASPIAN JOURNAL

MANAGEMENT AND HIGH TECHNOLOGIES

FORECASTING METHOD OF MUNICIPAL UNITS’ SOCIO-ECONOMIC INDICATORS BASED ON EVOLUTIONAL CONNECTIONISTS MODELS

Read Al-Qataberi Aiman, Kamaev Valery A., Shcherbakov Maxim V.  FORECASTING METHOD OF MUNICIPAL UNITS’ SOCIO-ECONOMIC INDICATORS BASED ON EVOLUTIONAL CONNECTIONISTS MODELS // Caspian journal : management and high technologies. — 2011. — №4. — pp. 40-45.

Al-Qataberi Aiman - Post-graduate student, Volgograd State Technical University, 28 Lenin avenue, Volgograd, 400131, Russia, aiman.qataberi@gmail.com.

Kamaev Valery A. - D.Sc. in Techology, Volgograd State Technical University, 28 Lenin avenue, Volgograd, 400131, Russia, cad@vstu.ru.

Shcherbakov Maxim V. - Cand. in Technics, Volgograd State Technical University, 28 Lenin avenue, Volgograd, 400131, Russia, maxim.shcherbakov@gmail.com.

The article considers the problem of socio-economic indicators forecasting in domain of the municipalities network in the Volgograd region. The birth-rate has been chosen as an indicator in case study. The information cards contain 29 sections of data on municipal units in the period 2007–2009. The peculiarities of the domain are listed. A method for forecasting of indicators is suggested and it is based on hybrid forecasting models. For each indicator 2 models are created. The first model is generated for forecasting variables of one dimension in the considered node as a dependant variable out of dimensions in the same node. The second one is for definition of dependences between considered dimension and dimensions in other nodes but of the same type. The forecasting models are based on Kasabov’s neural networks evolving connectionists models. The structure of neural networks evolving connectionists models is formed during the learning process with adjustable weights. This approach is able to adapt to changes in the structure or functioning of the system and its environment. The results of birth rate forecasting in 4 municipalities of Dubovsky district in the Volgograd region are shown. The average forecast error for the hybrid model was 2,88 %. Conclusions about further research are presented in the article.

Key words: socio-economic indicators,forecasting,evolutional connectionist systems,hybrid models