模糊模式识别在成矿预测中的应用

Metallogenic Prognosis Using Fuzzy Pattern Recognition

本文在模糊模式识别的理论基础上,提出了矿床预测的一种数学方法.基于F-PFS法,在特征因素和权重的调整下求出最佳分类.聚类中心即为标准模式.按单因素判对率作为该因素权重(逆距离权重),依标准模式模糊向量与已知单元模糊向量之间确定关系产生不等式方程组求解各因素的权重(距离权重),从而实现单元的贴近度识别.从鄂东南地区铜及多金属矿的预测表明,这类方法是有效的.

On the basis of fuzzy pattern recognition, a kind of mathematic methods for metallogenic prognosisare proposed in this present paper. Under the adjustment of characteristic factors and weights and byusing F-PFS method an optimum classification may be obtained. The cluster certre is considered as astandard pattern. According to the rate of right judgement of the single factor, which taken as the complement weight of the distance of the factor, and on the basis of the relationship determined betweenthe fuzzy vector of the standard pattern and that of the known cells, a system of linear algebraic inequality could be established and solved for distance weights of different factors. Thus the recognition ofunknown cells is realized by proximity degree. These methods have been used in the prognostication ofcopper and multinetal ores in Southeast Hubei areas and were proved to be effective for this end.