Research Of Complex Geophysical Prospecting Modle For Exploration Of Multi-metal Mine In Area Of Baogedewula

With the deepening development of prospecting work, the main target of prospecting has beenturning to deep and concealed or blind orebody from the surface. Multidisciplinary complementary andintegrated prospecting mechanism is an inevitable requirement for prospecting, in exploration of mineralresources How to take advantage of existing exploration to search for concealed mineral deposits quicklyand efficiently is the key topic of the study.This researched topic relies on <Complex Methods Prospecting in Inner Mongolia AutonomousRegion New Barag flag Baogedewula area>and <Ready Exploration in Inner Mongolia AutonomousRegion New Barag flag Baogedewula area> two project, the latter is a continuation of the former. AsQuaternary overburden distributed thicker, wider, so geological prospecting flag was obvious. Throughyears of prospecting practice in the study area, it had made more effective prospecting along the mainline of the application of geophysical methods, combining the different stages of geological results.Ultimately, the geological-geophysical prospecting model has been summarized for looking for theepithermal polymetallic ore in this area.With the development of geological theory, we have been trying to develop research of theintegrated information prospecting theory. Variety of mineral resources prediction model had beenestablished, mostly based on statistical theory, not only require a massive training samples, and theprediction accuracy is not enough. In response to these issues, this topic will combine traditionalprospecting methods and machine learning technologies, the support vector classification is applied tothe model of geophysical anomaly for differentiating between ore and non-ore, its unique theoreticalfoundation (promotion nature of the sector, the structural risk minimization, kernel function, the optimalseparating hyperplane), can solve effectively previous prediction models which require large samples,overfitting, local minima, the problem of high dimension. Currently, Support vector machines in manyfields has achieved successful application, and has been recognized by the majority of scholars.The research of the test model is built on the basis of above prospecting model, make use of Lead,zinc, silver polymetallic mine as discriminant object, carry on the research which based on supportvector machine of geophysical anomalies for differentiating between ore and non-ore. In the study of alarge number of domestic and international application examples in related fields, analysis the premise ofmetallogenic prediction of geology-geophysics, elaborated on the content and framework of statisticallearning theory and support vector machine. Completed correlation analysis of ore-body grade andgeophysical parameters (resistivity, polarizability, magnetic, etc.), and selected strong correlationelements as a predictive model input feature vectors, established a model for distinguishing the property of underground rock and mineral property with different depths. Through large number of previousstudies and comprehensive comparative analysis, RBF has been selected as kernel function of theprediction model. Through cross-validation and grid search the best model parameters have beenobtained and a set of discriminant model was established which based on geophysical anomalies fordifferentiating between ore or non-ore. Using the Visual C++6.0language programming, andcompleted the classified differentiated system. Finally, the model has been applied to district oflead-zinc-silver point in BaoGedewula Hanggaiyinhundi for instance validation, and achieved goodresults.The main results of the research topic:①S ystematicallywe summarized combination of complexgeophysical methods for prospecting, and built up geological-geophysical model for prospecting theepithermal polymetallic ore in this area.②B ased on the prospecting model and support vector machinetheory, the classified discriminant quantized model was established which own to geophysical anomaliesfor differentiating between ore and non-ore.