基于IFA的组合分析算法及其在矿产预测中的应用

矿产预测是寻找潜在矿产资源、增加后备储量的主要手段。地球化学信息作为直接找矿信息,在矿产预测中占据重要地位,对地球化学数据进行矿致异常的识别和评价是关键环节之一。地球化学数据分析的目的是从杂乱无章的原始数据中寻找并有效圈定化探组合异常。针对目前常用于组合异常分析的R型聚类分析和因子分析等传统算法对数据源限制较多、预测效果不理想等问题,将信号处理领域的盲源分离算法——独立因子分析(IFA)引入到矿产预测中,提出了基于IFA的组合分析算法,并选取广东汕头澄海区鸿沟山锡金多金属矿区地球化学数据作为研究对象。验证结果表明,基于IFA的组合分析结果更有利于解释矿产分布规律,圈定的化探组合异常更符合实际勘查结果。依据分析结果进行矿产预测,成功圈定出3个找矿靶区。

IFA-based combination analysis algorithm and its application in mineral prediction

Mineral prediction is the main means of discovering potential mineral resources and increasing backup reserves. Geochemical information, as direct prospecting information, plays an important role in mineral prediction. It is one of the key links to identify and evaluate mine-induced anomalies by geochemical data. The purpose of geochemical data analysis is to find and effectively delineate geochemical combination anomalies from disorganized raw data. In view of traditional portfolio analysis algorithms such as R-type cluster analysis and factor analysis with data source limitations and low accuracy, we introduce independent factor analysis (IFA), a blind source separation algorithm in signal processing, into mineral prediction, and propose a combination analysis algorithm based on IFA. Then we select the geochemical data of the Honggoushan Sn-Au polymetallic mining area in Chenghai District, Shantou City, Guangdong Province as the research object. The experimental results show that the results of the IFA-based combination analysis are more conducive to explain the distribution law of minerals, and the delineated geochemical exploration anomalies are more in line with the actual exploration results. According to the analysis results, we conducted mineral prediction and successfully delineated three prospecting targets.