成矿预测数据统计方法

利用统计方法对已有数据进行处理，进而推断其内部规律，并据此进行隐伏矿体定位仍是成矿预测的主流。从数学理论分析角度，讨论多元统计、地质统计分析、分形和随机过程等多种方法的数学表述及其物理内涵，并分析了这些方法及其分支的内在联系。针对数学方法的特点与地质数据的属性，提出了Weibull分布模型、自仿射分形和随机点过程分析是成矿预测领域较有应用前景的方法；多种传统方法相互配合使用，不但有利于数据集的深度挖掘，还会促进更有效的数学工具的开发。给出了Weibull分布与分形模型的配合使用有效描述胶东矿集区大尹格庄金矿成矿元素分布的应用实例。

Statistical Methods on Metallogenic Prognosis Data Processing

Processing the data obtained by statistical method and deducing their intrinsic regulations an important approach for the locaton of concealed ore body. From the point of view of mathematic principle analysis, the authors discussed the mathematic principles and physical meanings of multivariate statistical analysis, geological statistical analysis, fractal and random process, and also illuminated the relation among the different methods. Based on the characteristics of mathematic methods and the attributes of geological data, it was suggested that self-affine fractal and random process analysis would have an expansive foreground in metallogenic prognosis. The joint application of the traditional mathematic methods facilitates the data analysis and the method improvement. For instance, the assembled employment of Weibull model and the fractal succeeded in describing the distribution of the ore-forming elements of the Dayingezhuang ore deposit, Jiaodong Peninsula.