A method and device for constructing a geographically weighted regression model for mineral exploration

Abstract

The invention provides a method and device for constructing a geographically weighted regression model for mineral exploration. A plurality of elliptical screening areas are constructed by selecting target sample points in the target work area, and the attributes included in the target sample points are calculated using the weighted least square method. The characteristics and the corresponding mineralization are fitted to obtain multiple undetermined geographic weighted regression models corresponding to the target sample points, and then the undetermined geographic weighted regression model corresponding to the smallest residual square sum is selected as the optimal geographic weighted regression model. In this way, taking mineral exploration based on the geochemical data of river sediments as an example, the influence of the anisotropic characteristics of various geological elements related to mineralization can be characterized, and the space of complex geological processes and their products can be objectively reflected. The distribution is conducive to improving the analysis accuracy of mineral exploration.

Description

technical field [0001] The invention relates to events and processes with anisotropic characteristics in the technical field of mineral exploration, in particular to a method and device for constructing a geographically weighted regression model for mineral exploration. Background technique [0002] In the process of mineralization and mineral resource exploration, the Geographically Weighted Regression (GWR) analysis model (Geographically Weighted Regression, GWR) is often used to perform local weighted least squares regression on a small number of samples in a given area to estimate the local relationship between geographical variables. Reveal detailed changes in localized areas. [0003] Taking metallogenetic geological events as an example, due to the nonlinear characteristics of the metallogenic process, such as complexity, anisotropy, irreversibility, and self-similarity, the traditional geographically weighted regression model believes that sampling points at equal di...

AI Technical Summary

Problems solved by technology

[0003] Taking metallogenetic geological events as an example, due to the nonlinear characteristics of the metallogenic process, such as complexity, anisotropy, irreversibility, and self-similarity, the traditional geographically weighted regression model believes that sampling points at equal distances from the regression point have an important influence on the regression point. With the same influence, it cannot objectively reflect the complex geological process and the spatial distribution of its products, and cannot consider the anisotropy characteristics of various geological elements related to ore-forming. The analysis accuracy is low, which affects the accuracy of mineral exploration

Images