ARTIFICIAL NEURAL NETWORK APPLICATIONS IN AQUACULTURE AND FISHERIES RESEARCH AWARENESS ON SELF-ORGANIZING MAP

摘要

Self-organizing maps (SOM) is a clustering tool invented by a Finnish academician, Teuvo Kohonen in 1981. SOM provides a way of representing multidimensional data in much lower dimensional spaces that makes better visualisations of the existing, if there are any, relations between parameters,According to Gene et al (2011), Kohonen's SOM are producing and preserving topological relations between every parameter of the input vectors (Kohonen 1995). Unlike multilayered feed forward neural networks that are the most common neural network modelsused in large variety of applications such as engineering and economy problems, SOM employs unsupervised learning training mechanism. Brosse et al (2001) stated that SOM typically displays a high dimensional data set in a lower dimensional space.