Geochemical Abnormity Identification Based On Kernel Fisher Discriminant Analysis

In the study of mineral exploration about geochemistry, the most important process in evaluating the geochemical anomalies is to identify the geochemical anomalies caused by ore induced abnormity or not.to ring-fence the reasonable prospecting area Finding the delineation of the abnormal area effectively is not only useful in looking for mine, but also can reduce the material, save money and reduce human resources. Therefore, the focus of this paper is to identify the geochemical anomalies.In the data processing, because of the most factors about geochemistry exist In the research field of regional geology in the form of affected or interrelated with each other, it may lead to the non-mining and mining-induced abnormalities in the delineated abnormal area Fisher as a discriminant method on the study of the geochemical anomalies identification, we can’t just consider few factors,we should take indicators into account as many as possible, and then analyze the relationship between these indicators, these steps will greatly enhance the accuracy in the abnormality recognition.Fisher is an effective method to identify abnormity, which selects the best projective vector and then project on the geochemistry data, at last it will divide the data into different kinds. The idea can provide data support for the delineation of ore induced anomaly However, due to the complexity of the system of geology, the process of the mineralization and other elements’ enrichment is complex, and with different age of the mineralization it makes the geochemistry abnormity fuzzily. The linear discriminant function can’t well represent the relationship on the basis of the fisher discriminant analysis Therefore, it is necessary to introduce the kernel function for determining the geochemical data.The method of kernel refers to the input the inseparable spatial data by using a nonlinear mapping to map these data to a high dimensional or infinite-dimensional feature space, in the end, the spatial data become separable data.so we can take linear analysis method for feature extraction in the feature space. Thus, the Fisher discriminant analysis method is transformed into a much more efficient Kernel Fisher methods to identify geochemical abnormality.This paper studies the use of kernel Fisher discriminant analysis in geochemical abnormity identification. The idea of kernel Fisher discriminant analysis to classify abnormal sentenced principle refers to with the "core skills" to input data space converted into a nonlinear feature space implicitly, so we can use the linear Fisher discriminant analysis to identify the data abnormity in the transformed space On the research of 1:20 000 stream geochemical data of southeast Hubei, it indicates the kernel Fisher analysis which is based on geochemical anomaly recognition is remarkable.