Development of an expert system for echogram analysis in order to identify the species of pelagic schools <B>Development of an expert system for echogram analysis in order to identify the species of pelagic schools</B><P> S. Georgakarakos, G. Charalampous G. Paterakis and N. Tsimenides</B><I><P> Institute Marine Biology of Crete</I><B><P> Abstract</B><P> Georgakarakos, S. <I>et al </I>1993. Development of an expert system for echogram analysis in order to identity the species of pelagic schools. Proc. 4th Natl. Symp. Oceanogr. Fish., pp. 314-317.<P> IMBC has developed an expert system, which is able to extract morphological, energetic and spatial-time parameters of pelagic fish schools from the acoustic signal and also to use these parameters as a guide to train a multi-layered food-forward neural network. The expertise of the trained network was tested by inputting cases of known species and comparing the predicted species with the original ones (:given by trawling). Finally the trained network was used to suggest the species identity of the rest of 1224 schools. These preliminary results indicate that it is possible to identify the species using the hydroacoustic parameters describing a fish school.<P> Proceedings of the 4th National Symposium on Oceanography and Fisheries, Rhodes isl, (Greece), 26-29 April 1993, pp. 314-317<P>