eCite Digital Repository

Multielement geochemical data missing value and below detection limit imputation

Citation

Cracknell, MJ and Baker, MJ and Cooke, DR and Escolme, A, Multielement geochemical data missing value and below detection limit imputation, SEG 2019 Conference, 7-10 October 2019, Santiago, Chile, pp. 1-2. (2019) [Conference Extract]


Preview
PDF (Published Abstract)
Pending copyright assessment - Request a copy
205Kb
  

Official URL: https://www.segweb.org/SEG/Events/SEG_Conference_A...

Abstract

Introduction

The precise measurement of small concentrations of elements from chemical analyses is limited due to instrument sensitivity. Such concentrations are known as the detection limit (DL). Given the increasing popularity of multivariate data analysis tools to identify patterns within geochemical data, which often require all data to be represented by real values, methods for imputing missing or below DL (BDL) values are required. BDL values are often either removed completely or replaced with a constant value such as half the BDL (Carranza, 2011; Hron and others, 2010). Where additional information is available, such as with multivariate data, BDL values can be imputed using methods that assess the conditional similarities between samples within data space.

Item Details

Item Type:Conference Extract
Keywords:geochemistry, mineral chemistry, LA-ICP-MS, imputation, expectation maximisation
Research Division:Earth Sciences
Research Group:Geochemistry
Research Field:Exploration Geochemistry
Objective Division:Mineral Resources (excl. Energy Resources)
Objective Group:Mineral Exploration
Objective Field:Copper Ore Exploration
UTAS Author:Cracknell, MJ (Dr Matthew Cracknell)
UTAS Author:Baker, MJ (Dr Michael Baker)
UTAS Author:Cooke, DR (Professor David Cooke)
UTAS Author:Escolme, A (Dr Angela Escolme)
ID Code:137323
Year Published:2019
Deposited By:Earth Sciences
Deposited On:2020-02-10
Last Modified:2020-03-06
Downloads:3 View Download Statistics

Repository Staff Only: item control page