Abstract
This paper proposes the use of concepts from fuzzy-set theory applied to the factor loadings from principal components analysis as a method to derive innovative regionalizations of geographic flow data. The method is developed in the context of analyzing US interstate migration patterns. As in previous literature on the topic, migration regions are identified as distinguishable clusters of geographic subareas that share similar migration 'profiles', that is, similar distributions of outflows to alternative destinations or inflows from alternative origins. A fuzzy-set perspective allows geographic subareas to be regarded as belonging with varying degrees of intensity to a migration region. The relative level of 'belongingness' can range from subareas in the migration region's core, where fuzzy information units or 'fit' values approach 1, to subareas external to the migration region where they are zero. In between are peripheral sections of the migration region with intermediate fits. Thus, migration (and other similar geographic) regions are represented as non-mutually exclusive entities having fuzzy rather than sharp boundaries. US migration regions for 1980-1981 to 1991-1992 are examined and shown as dot-density maps. Measures are proposed for assessing the relative fuzziness of migration regions, the geographic extent of their overlaps and the persistence over time of their regional structure.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 141-162 |
| Number of pages | 22 |
| Journal | Geographical and Environmental Modelling |
| Volume | 2 |
| Issue number | 2 |
| State | Published - 1998 |
Keywords
- Demographic method
- Internal migration
- Methodology
- Migration
- United States
ASJC Scopus subject areas
- Geography, Planning and Development
- General Earth and Planetary Sciences