A model of cluster and discrimination for fisheries science

The present paper is an introduction of a universal model of cluster and discrimination in the field of fisheries science. The mathematical principle and algorithm are described and the application results analysed. This model is called “weighted generalized distance” method, based on the principle of “minimum distance”. For different types of data, different methods are put forward. The quantitative attributes are standarized by “norm” method and the qualitative attributes are quantified by indicative value based on Shannon's information theory to locate the specimen in the coordinates. An approach by progressive readjustment of pivotal points is employed for the cluster rational. Finally, any new specimen can be discriminated by the principle of “minimum distance”, too.