The article considers the main construction principles and an architecture of the unified platform for data mining. The methodology for organizing output of recommendations on estimation of organization financial management indices over a number of reporting periods on the expert level is investigated. The methodology is based on the estimation of trends characters of the corresponding indices and the inference of linguistic values of corresponding objects properties of the applied ontology. The methodology also allows to construct the expert opinions of both current state of organization indices and forecasting one obtained by the application of the F-transform algorithm. The object of the analysis is a commercial organization. The modeling process is the output of expert opinion on the financial management of the object under consideration.