Kopulalarla Maden Kaynak Kestirimi

Abstract

This thesis basically includes mineral resource estimation for this purpose, kriging is method widely used in mining industry. But kriging has some disadvantages such as non-resistance of variogram to outlier values, smoothing of kriged estimate, unrealistic assumptions in the assessing uncertainty of estimated values. In this thesis, copula is suggested to avoid these problematic issues. Copulas are functions that exhibit relationship between for which marginal distributions are uniform. In this thesis, distance based copula functions are developed, effect of the number of neighboring data and nugget effect of the estimation is studied. The estimation of the mean value of the blocks by copulas is also investigated. Two case studies for estimation with copulas are presented for normally and log-normally distributed data set.