Applicaiton Research Based On NMF About Geosciences Maps Information Integration And Mineral Resources Prediction

The thesis systematically describes the information integration of geosciencesgallery and the application of mineral resources prediction. Geosciences galleryinformation integration refers to the integration of data, because it features multiplecollection methods, complex data types, wide span of the time and space of the datadescribed the phenomena. Sorting the built database of geosciences data, sharing thedata and improving the utilization of geosciences data in China is of great significanceto boost the development of geosciences research.Mineral resources prediction, namely, the mineral resources potential assessment,is to use the integrated information of geosciences gallery to estimate the mineralresources potential and assess its value for further deployment of geological work,which is quite important to the strategic development of geological work. With themineral resources prediction becoming more and more difficult as well as computertechnology continuing to evolve, the proper use of computer technology on carryingout data mining and mineral resources prediction may greatly improve the efficiencyand accuracy of the prediction work.This thesis introduces the basic principles and the nature of the algorithm of NMF(Non-Negative Matrix factorization) from the perspective of information integration,and the author adds two constraint conditions to the objective function of thetraditional NMF methods and applies iterative rules to improve the algorithm. Finally,the author adopts the improved CNMF algorithm for quantitative prediction on thesilver mineral resources of eastern Inner Mongolia at1:200,000, and compares thepredictable results with the results obtained via traditional NMF algorithm andverifies the reliability of the algorithm of information integration of geosciencesgallery.