Fuzzy Regression in Fisheries Science: Some Methods and Applications
S.B. Saila and S. Ferson
- Price: $0.75 Sale: $0.00
Methodology for fuzzy regression is described and then illustrated with some applications to published data from the fisheries literature. Fuzzy regression should be a useful alternative or complement to conventional statistical regression whenever the relationship between variables is imprecise, data are imprecise, or sample sizes are very small. Formulations of fuzzy regression models are described, and goodness-of-fit criteria are briefly investigated and compared. The fuzzy regression technique is first illustrated by applications to data from the relationship between the proportion of chinook salmon jacks entering a river and the proportion of jacks in the ensuing cohort. A second example showing the relation between chinook salmon yearlings and water flow is also investigated. The third example deals with the relationship between an abundance index and virtual population analysis results for a short-time series of Atlantic mackerel data. Fuzzy regression methodology seems relatively straightforward, but it usually involves a constrained minimization problem, which may require mathematical programming methods. On the basis of this review and applications, we believe that fuzzy regression has utility in some fisheries-related applications, but further evaluation is suggested.
- Item number: AK-SG-98-01p
- Year: 1998
- DOI: https://doi.org/10.4027/fsam.1998.16