A case study on geochemical anomaly identification through principal components analysis supplementary projection |
| |
| Institution: | 1. CVRM—Geosystems Center, IST, Av. Rovisco Pais, 1049-001 Lisboa, Portugal;2. Earth Sciences Dept., Univ. Coimbra, Lg. Mq. Pombal, 3000-272 Coimbra, Portugal;1. Dept. of Surgery, Western University, London, ON, Canada;2. Division of Transplantation, Dept. of Surgery, Rush University Medical Center, Chicago, IL, United States;3. Cantonal Hospital Winterthur, Winterthur, Canton Zurich, Switzerland;4. Dept. of General and Visceral Surgery, Frankfurt/Main, Germany;1. Department of Surgery A, Hospitais da Universidade de Coimbra, Centro Hospitalar e Universitário de Coimbra, Coimbra, Portugal;2. Clínica Universitária de Cirurgia III, Faculty of Medicine, University of Coimbra, Coimbra, Portugal;3. Department of Pathology, Hospitais da Universidade de Coimbra, Centro Hospitalar e Universitário de Coimbra, Coimbra, Portugal;1. Department of Statistics, Ensai, Bruz, France;2. Agrocampus-Ouest, Rennes, France;3. Département de Mathématiques et de Statistique, Université de Montréal, Montréal, Canada;4. Applied Mathematics department, Agrocampus Ouest, Rennes, France;1. Department of Urology, University of Bologna, S Orsola-Malpighi Hospital, Bologna, Italy;2. Department of Pathology, University of Bologna, S Orsola-Malpighi Hospital, Bologna, Italy |
| |
| Abstract: | Based on a case study in which a single geochemical anomaly was located in the vicinity of an abandoned mine in Central Portugal, a recursive methodology for anomaly/background separation was developed. This methodology relies on the supplementary projection of each of the samples taken from a subset of ‘anomaly candidates’ onto the axes provided by Principal Components Analysis of the background subset. The concept of ‘anomaly intensity’, defined by the average of the distances from the original to the supplementary projections, is the basis for final anomaly identification. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
