Research And Application Of Support Vector Machine Classification Method Based On Ant Colony Algorithm

Geochemical data has been a key source of information for mineral exploration.With the progress and development of the times,the processing of geochemical data has changed from a single traditional method to a combination of modern mathematical methods.the goal of geochemical data processing is to extract useful information from raw geochemical data through mathematical methods from a large number of observational data and complex conditions,revealing the intrinsic connections and phenomena between indicator variables(mainly chemical elements)and various geology.In order to find methods and techniques for exploring geochemistry and providing exploration for mineral resources and solving geological problems.So far,from the data processing point of view,the data classification problem is one of the research hot spots,the algorithm of different data characteristics will naturally be different.The data classification ability of support vector machine is ideal,but the optimal parameter value will often be different,and the result will be different.Therefore,based on the support vector machine algorithm,this paper proposes an ant colony support vector machine classification,both classification accuracy and convergence speed have been greatly improved.This paper first introduces the basic principle of support vector machine,then introduces the principle of basic ant colony algorithm.The ant colony algorithm and support vector are analyzed first advantages and disadvantages of the machine,and put forward the combination strategy of the two algorithms.In this paper,support vector machine and ant colony optimization support vector machine are compared to verify the feasibility of this algorithm.It is proved that the new algorithm combined with ant colony and support vector machine is feasible and has obvious advantages,and the complexity of the algorithm is reduced to a certain extent.Then the empirical analysis of this algorithm is carried out.In this paper,the relationship between the spatial distribution characteristics of geochemicalelement content data of sediments in Miraman porphyry copper polymetallic mining area is analyzed.Finally,the ant colony optimization support vector machine(SVM)algorithm is applied to 1:20 in Miraman area of Tibet ten thousand water system sampling geochemical data.Firstly,the attribute model of ore-forming data is constructed by verifying the data set with or without ore-spot label.Then,the classification and prediction of ore-mining sites were made by using ant colony optimization support vector machine.The ant colony support vector machine algorithm combines the advantages of the two algorithms and achieves 83.1%accuracy in the test set.The marking of ore-site is realized.