The use of Procrustes Target Analysis to discriminate dominant source regions of fine sulfur in the western U.S.A. |
| |
| Institution: | 1. Department of Obstetrics and Gynecology, University of Pennsylvania, Philadelphia, Pennsylvania;2. Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden;3. Department of Internal Medicine, Division of Endocrinology and Metabolism, Hacettepe University School of Medicine, Hacettepe, Ankara, Turkey;4. Department of Obstetrics and Gyneocology, Reproductive Medical Center, Peking University Third Hosptal, Bejing, People''s Republic of China;5. PCOS Challenge, Atlanta, Georgia;6. Gynaecworld, Mumbai, India;7. Department of Psychiatry, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, Pennsylvania;8. Monash Centre for Health Research and Implementation, School of Public Health and Preventive Medicine, Monash University and Monash Health, Clayton, Victoria, Australia |
| |
| Abstract: | Receptor modeling, using Procrustes Target Analysis (PTA), was applied to a fine sulfur data set to discriminate dominant sources and areas of influence in the western U.S.A. The work presented in this note is closely linked to a paper published by Malm et al. (Atmospheric Environment24, 3047–3060, 1990), which used two different analyses, area of influence analysis (AIA) and rotated principal component analysis (RPCA), to show similar areas of influence for most source regions in the western U.S.A. A critical examination of RPCA revealed that two issues might confound interpretations. These are (i) eigenvalue degeneracy combined with trying to fit nine orthogonal dimensions pairwise in the PC space and (ii) statistical issues related to missing data and eigenvector truncation. In determining how much of the AIA results are present in the data spanned by the PC space, our findings indicated a 164% improvement in the variance overlap when PTA was substituted for RPCA. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
