A Classification Method for Anomalous Chemical Single Elements with Constrained SVM

Abstract

The invention provides a method for classifying abnormal chemical single elements with constrained SVM. The method includes the steps of: preprocessing the sampling data within the range to be classified to obtain a training data set; adding constraints to the training data set to obtain a training data set containing constraints ; Normalize the training data set containing the constraints, and the normalized training data set containing the constraints is used to train the SVM classifier; obtain the optimal kernel function of the SVM classifier to obtain the SVM optimal classifier; Use the SVM optimal classifier to linearly classify the training data set; use the results of the linear classification to generate a complete geochemical single-element anomaly map. The invention effectively delineates geochemical single-element anomalies that meet the constraints of geological conditions, and effectively solves the problem of distribution and direction of single-element anomalies.

Description

technical field [0001] The invention relates to the field of geochemical prospecting thematic mapping, in particular to a method for classifying anomalous chemical single elements with constrained SVM. Background technique [0002] The single-element geochemical anomaly map shows the chemical anomaly characteristics and scope of a single element (indicator), which is an important achievement map in geochemical exploration, scientific research and production, and is also the basic map for obtaining a comprehensive anomaly map for anomaly classification and evaluation. At present, computer-aided mapping of single-element anomalies is generally carried out by tracking contour maps based on regular grids or irregular triangulations. The generated anomaly maps are completely dependent on the sampling data and are greatly affected by the uniformity of sampling. [0003] According to ore deposit theory and expert experience, the shape of anomalies should correspond to geological fe...

AI Technical Summary

Problems solved by technology

[0004] Existing methods based on grid-based contour tracing to delineate single-element anomalies seldom consider the constraints of geological conditions. In practical applications, geochemical prospectors often consider geological Conditions, but it is greatly affected by human beings. It is necessary to explore a method that uses intelligence to delineate single-element anomalies based on expert experience, and then quickly delineate combined anomalies.

Images