Identifying regular blocks in valued networks: A heuristic applied to the St. Marks carbon flow data, and international trade in cereal products
Author
Summary, in English
While the concept of regular equivalence is equally applicable to dichotomous as well as valued networks, the identification of regular blocks in regular blockmodels is somewhat problematic when dealing with valued networks. Applying the standard procedure for identifying ties in such blockmodels, a procedure perhaps most suited for dichotomous networks, does tend to generate block images and reduced graphs that differ from intuitive notions of such structures.
This paper outlines a formal heuristic procedure for identifying regular ties in valued networks where the “significance” of ties is related to each actor's role sets. Combined with measures for block criteria fulfillment, the procedure yields reduced graphs, which seem more sensitive to patterns, rather than strengths, of ties.
Two data sets are used as examples in this paper: the St. Marks carbon flow web dataset, and a new dataset containing international trade flows of cereals and cereal products based on Comtrade data.
This paper outlines a formal heuristic procedure for identifying regular ties in valued networks where the “significance” of ties is related to each actor's role sets. Combined with measures for block criteria fulfillment, the procedure yields reduced graphs, which seem more sensitive to patterns, rather than strengths, of ties.
Two data sets are used as examples in this paper: the St. Marks carbon flow web dataset, and a new dataset containing international trade flows of cereals and cereal products based on Comtrade data.
Department/s
Publishing year
2007
Language
English
Pages
59-69
Publication/Series
Social Networks
Volume
29
Issue
1
Full text
- Available as PDF - 309 kB
- Download statistics
Document type
Journal article
Publisher
Elsevier
Topic
- Social and Economic Geography
Keywords
- Network analysis
- Regular blockmodels
- Valued networks
- Regular blocks
Status
Published
ISBN/ISSN/Other
- ISSN: 0378-8733