基于逻辑回归模型的化探异常信息识别研究——以克拉玛依地区为例

化探数据对矿产资源勘查工作有着重要作用,其中比较关键的工作就是从化探数据中识别矿床相关的化探异常信息.在化探异常信息识别工作中也发展出了较多的技术,但是它们大多针对单变量进行分析.为了对多变量化探数据进行分析并识别金矿相关的地球化学异常信息,笔者将逻辑回归模型用于研究区化探数据分析,通过研究区内对金矿预测比较有价值的16种元素的逻辑回归建模及模型应用,发现逻辑回归是一种有效的化探多变量数据分析和建模技术,研究结果显示,笔者建立的逻辑回归模型不仅可以有效识别已知金矿区的地球化学异常信息,而且对那些还未发现矿床的区域具有预测作用,能够为矿产资源勘查工作重点区域的选择提供指导.

Application of Logistic Regression Methods in Geochemical Data Analysis and Mineral Exploration: Example from Karamay Region

Geochemical data is essential for mineral exploration, and one of the main challenges is how to identify the anomaly that was related to the formation or locations of mineral deposits. Many techniques have been developed to identify geochemical anomalies in the past years, but most of these techniques are designed for univariate data. To identify geochemical anomalies from multivariate geochemical data and to get gold deposits related information, logistic regression method is used to analyze geochemical data (sixteen hydrothermal/epithermal elements are included) of this study area. The results demonstrate that the developed logistic regression model is effective for geochemical anomalies identification and gold prediction, because the model can not only identify the geochemical anomalies where there are known gold deposits, but also identify other strong geochemical anomalies where there is no known deposit. Therefore, the logistic regression method is recommended to be used to geochemical anomalies identification and mineral prediction.