Multivariate geochemical anomaly identification method based on graph attention self-encoding

Abstract

The invention discloses a multivariate geochemical anomaly recognition method based on graph attention self-encoding, comprising the following steps: determination of composition threshold K; construction of topological network of geochemical elements; learning of multivariate geochemical features; background reconstruction of multivariate geochemical elements; value calculation. The present invention introduces graph learning into geochemical anomaly detection, uses K nearest neighbors to construct a topological relationship graph of geochemical elements, and constructs and trains a graph attention self-encoder capable of simultaneously extracting element composition relationships and spatial structure features. Based on the trained graph attention self-encoder, the multivariate geochemical background is reconstructed, and the final outlier value of each sampling point is calculated. The present invention expands the existing neural network model, so that the model can directly process sampling point data and be applied to irregular areas, greatly improves the learning performance of the background of chemical exploration and the practicability of the model, and provides a practical solution for complex geological conditions. Reliable method for identifying anomalies in geochemical exploration.

Description

technical field [0001] The invention relates to the field of multivariate geochemical exploration anomaly identification and the field of artificial intelligence application, in particular to a multivariate geochemical anomaly identification method based on graph attention self-encoding. Background technique [0002] The identification of multivariate geochemical anomalies is one of the important contents of mineral resources exploration, and its anomaly information helps geologists to judge potential mineral deposits. Geochemical anomalies have strong spatial heterogeneity, and the spatial heterogeneity of geochemical elements must be considered. Traditional anomaly identification methods such as fractal / multiple analysis, kriging method, and spatial factor analysis take into account the correlation of spatial neighbor samples, and are outstanding in the identification of geochemical anomalies. In recent years, with its good and automatic learning ability for complex spati...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the problem that the existing convolutional self-encoder cannot directly use sampling point data and fails to take into account real sampling points and irregular ore-forming areas, the present invention provides a multivariate geochemical anomaly identification method based on graph attention self-encoding, Include the following steps:

Images