| Objection of exploring mine is turning from surface ore shallow ore easy to identify ore to lie low ore deep ore difficulty to identify ore, with its increasing transformation, finding a kind of effective method of located prediction of lie low ore has becoming a frontier and hot point of mining and mining prediction. At the same time with the rapid development of modern computer technology, description mining model(DMM) which is expressed by nature language has not meet need of modern mining prediction, In order to adapt to new situation and increase efficiency, it is imperative that conventional method is transformed. Both fully exhibit the function of traditional DMM and directly apply it to located prediction of modern mine, we must explore fully new expression manner of mine bed, so that not only take advantage of useful information of DMM but realize linking up with high and new technology of located prediction of modern resource. Therefore, appraisement system of numeral model (ASNM)has become a kind of necessary trend of studying and developing of lie low ore location predictionASNM would transform DMM into data knowledge information and symbol that computer can directly identify and deal with, so that build a intelligent and inferential network to realize knowledgenumeral and intel1igentizing .Based on background of huize plumbum and zinc, this paper make use of neural network and fuzzy mathematic and build ASNM. This system includes 3 net models: SOFMANN is characteristic of self-adapting, self-organizing, learning without teacher control mineral multi-information; FCA can take full advantage of experts and fuzzy mathematics processing fuzzy phenomenon and fuzzy action in concealed mineral prediction; BP model is a kind of monitoringtraining with teacher information, it has some generalizing ability, this ASNM system has accomplished location prediction of concealed ore, total efficiency is more than 70 percent... |
PDF Full Text Request