THE FUZZY GROUP METHOD OF DATA HANDLING WITH FUZZY INPUT VARIABLES

The problem of constructing forecasting models with incomplete and fuzzy input data is considered in this paper. For its solution Fuzzy Group Methods of Data Handing (FGMDH) with fuzzy inputs is suggested. The method enables to construct a forecasting fuzzy model using experimental data which are not distinct. The method was implemented as software kit and experimental investigations of were carried out in the problem forecasting stock-prices at the Russian stock-exchange. The comparison of the suggested method with known methods: GMDH and fuzzy GMDH are also presented.