QA204 : Spatial Outlier Detection
Thesis > Central Library of Shahrood University > Mathematical Sciences > MSc > 2013
Authors:
Maryam Moghaddam [Author], Davood Shahsavani[Supervisor], Mohammad Koneshloo[Supervisor], Hossein Baghishani[Advisor]
Abstarct: In ore deposits, detecting the grades that are in the right tail of the distribution are economically important. As a result, detecting the cutoff of these observations is important in mining industry extraction. Existing methods for detecting spatial outlier indentify the outlier values but do not provide criteria for identifying cutoff for high values. Hence, in this thesis, Top-Cut method is suggested to identify the cutoff for high values which is baxsed on absence of edge effect. In the absence of edge effect, highe values have no influence from lower grades. In this method, ratio of the indicator cross variogram and indicator variogram in each suggested cutoff is calculated and if this ratio is constant, in terms of function of distances, it shows absence of edge effect. Also if the variogram of the residual indicator for suggested cutoff shows pure nugget effect, then the suggested cutoff is chosen for topcut grade. Evaluation of these methods and some outlier detection methods has been carried out in a simulation study. Moreover as a real case study, this method has been used to determine the cutoff of high-grade of Sari-Gunay deposit in Kordestan province, Iran. We found that the value 4.63 ppm as the cutoff for high-grade of Sari-Gunay deposit.
Keywords:
#Spatial Outlier #high-grade #Edge effect #Indicator variogram #Indicator cross variogram Link
Keeping place: Central Library of Shahrood University
Visitor: