Abstract
Hitherto, the ranges of Zn-Pb orebodies have been determined on the base of Zn + Pb contents in boreholes. In general, this approach give good results, but not ever when concerning data from older drillings (frequent decrease of barren or ore minerals from drill cores). This problem could be solved by using not only contents ofmetals, but other parameters too. Because of numerous of such parameters, the methods of supervised classification i.e. the pattern recognition methods (the potential function and "Kora-J" algorithms) were used to performance the multidimensional data to one number only - the probability that particular boreholes would reveal the characteristic conditions for occurrence of orebody. To compute the probabilities in localities among existing drillings, the gcostatistical methods were used. Using the baycsian equation, these probabilities were changed into probabilities, that particular boreholes would recover an orebody. Investigations were conduct on the base of data originated from 273 boreholes, drilled in the mine area (before its foundation) and its vicinity. At first, cach drilling was described by 386 lithological, stratigraphical and tectonical features. Before classification, this number was reduced (Puri-Sen-Tamura algorithm) to 14 informative features. Though with shortness, obtained results allow one to indicate the areas of high probability of occurrence of orcbodics, Such areas, inside exhausted part of the mine, can be verified on the basis of the comparison with the borders of orcbodics. which arc proved by the exploitation. It is also possible to compare the areas with the orebodies, pointed in the standard prognosis, based on the metal content in the boreholes. The two comparisons show, that the quality of the pattern recognition method prognosis is similar to the standard prognosis. Moreover, the first prognosis delivers the new information to the second one. The information can be used as auxiliary one to better determine real distribution of orebodies. Besides, in the stage of searching for informative features, the most informative ones were found. They should reveal connections with genesis of the ore deposit, but, in this case, such interpretation was impossible.
Go to article