Geochemical Data Processing And Application Based On The KNN Algorithm And ISOMAP Algorithm Research

Geochemical exploration data processing is an important content of exploration geochemistry, different methods of data processing directly affects the geochemical prospecting effectiveness and efficiency. Geochemical exploration data processing is applied mathematics method and computer technology, found in the original data from geochemical exploration and extraction of available information, the chemical elements and the internal connection of various geological phenomena, provide the basis for geochemical prospecting. How to effectively extract the geochemical anomalies information science, and to rapidly and accurately from a large number of abnormal screening evaluation, to determine further ore-prospecting target area, is the key to geochemical prospecting work. Geochemical element content value is not confined to normal distribution or lognormal distribution, the characteristics of discontinuity, respectively, heterogeneity, diversity and random characteristics, namely, nonlinear characteristics. In the geochemical exploration data processing, for nonlinear characteristics will adopt the algorithm of nonlinear.Based on qaidam large had town in qinghai province south slope area as an example, using traditional statistical method, KNN algorithm, cluster analysis and principal component analysis, ISOMAP algorithm in the study area 1:10 000 soil geochemical survey data analysis processing, on the basis of the understanding of the study area geology background, delineating metallogenic prospect areas, work for the geological exploration work provides the next target. From geochemical anomaly evaluation and research content, the use of the modern mathematics method and the nonlinear analysis method of mining in geochemical data of metallogenic exception information. Through comparing with the traditional analysis method show that the KNN classification algorithm of the element in the geochemical exploration data storage anomaly has good recognition effect. Using principal component analysis(Cu, Au, zinc, As, Sb, Pb six elements into two groups, based on the abnormal selects the combination of two groups of elements. Use of ISOMAP six elements can be divided into two groups, the same framed as the combination of the two groups of elements. Is obtained by comparison with the combination of ISOMAP algorithm selects elements than PCA tagged anomaly area distribution and shape the rules.