Selecting statistical models and variable combinations for optimal classification using otolith microchemistry.

The article presents a study which determines statistical models and variable combinations for optimal classification of fish origin by using otolith microchemistry. The study uses otolith fingerprint data sets of various fish species from three coastal areas such as French lagoons, Portuguese estuaries and U.S. coastal nurseries to assess fish habitat discrimination. Result shows that accuracy has varied according to the model used and the number of elements considered.