The Study On Artifical Neural Network And It's Application In GeoScience

During the past decades,geologists have accumulated a large quantity geologicaldata,such as geophysics,geochemistry and remote sensing.Alllof data can help us findsome character of geological body or geological phenomenon in some extent.It hasbeen an importante work faced by the geologists to integrate these multi-orientinformation to reveal the essence of geoscience.The traditional method use multista-tistics to build assessing model and then implement taget forcast, because of comple-xity of whole process and sample maybe not achieve the requirement of statistics,it isa very difficult work.futhermore,we face a lot of difficulty in information integrationand quantification.The neuralnetwork is the enmulation of mankind's brain,it is acomputing method oriented in the the study of brain working mechanism.The neuralnetwork consists of many processing unit,it need not mathematical model and canexpress non-linear relationship through the study of sample data.the neural networkcan solve the problem such as large number data,data variety and complex minedeposit condition in mineral resource assessment.The study ability of neural networkis very strong and it has the correction ability to the data which have noise anderrors,soit is widely used in geoscience.In this paper,we discuss the basic theory and method of neural network,andexpatiate the program of ART,HOPFIELD,BP,RBF network and its application ingeoscience.