东昆仑东段水系沉积物测量数据处理中因子分析法的应用研究

Study and Application of Factor Analysis for Stream Sediments Data Processing in the East of Dongkunlun

因子分析是一种多元统计分析方法,已广泛应用于地球化学数据处理中。其计算基础是相关系数矩阵(或协方差矩阵),易受特异值影响,加之地球化学数据是成分数据,其闭合效应会影响数据的相关系数矩阵。因此,在进行因子分析前须对数据进行变换,以获得准确的结果。以东昆仑东段水系沉积物地球化学测量数据为例,分别对原始数据、Log变换后数据及ILR变换后数据进行常规因子分析和稳健因子分析,旨在优选一种在地球化学数据处理中效果较好的因子分析方法,明确东昆仑东段的主攻矿种,并进一步圈定找矿靶区。研究表明:对ILR变换后的数据进行稳健因子分析发现,各主因子与各成因类型矿床对应性最好。利用其因子得分异常圈定了找矿靶区7处,靶区内已发现了4处大型以上的金矿床和多处中型金、铁和多金属矿床,通过进一步工作,在靶区内新发现了一些找矿线索,已发现的矿床(点)与靶区的异常规模还不相称,进一步开展工作的潜力巨大。

Factor analysis is a multivariate statistical analysis method,which has been widely used in geochemical data processing.The basis of factor analysis is correlation coefficient matrix(or covariance matrix),which is very vulnerable to data outliers of any kind,and the compositional data closure effect.Therefore,factor analysis in geochemical data processing requires data transformation to obtain accurate results.The geochemical data of stream sediment survey in the east section of Dongkunlun are used to conduct classical factor analysis and robust factor analysis on the raw data,log transformed data and ILR transformed data respectively,aiming to optimize a factor analysis method with better effect in geochemical data processing,make clear the main ores in east section of Dongkunlun and delineate the prospecting target areas.The study results show that the robust factor analysis of the ILR transformed data shows that the main factors have the best correspondence with various genetic type deposits.Based on its factor scores abnormal,seven prospecting target areas are delineated.Four large-sized gold deposits and seven medium-size gold,iron and polymetallic deposits are found in the above prospecting target areas,by further working,some prospecting clues newly obtained in the above areas.In general,the discovered deposits(ore points)are not commensurate with the abnormal size of the above prospecting target areas,and the potential of the further work in the areas is great.