Artificial Neural Networks (a new statistical approach) method in length-weight relationships of Alburnus Mossulensis in Murat River (Palu-Elaziğ) Turkey

Artificial Neural Networks (ANNs) is a reliable and alternative method used specially in the growth characteristics of some fish species. The purpose of this work was to compare with length-weight relationships and Artificial Neural Networks, the length-weight relation of 415 (232 female and 183 male) Alburnus mossulensis captured from Murat River between January 2018 and December 2018. The total length and the weight of the fish changed between 7.2-19.9 cm (12.81±0.095) and 3.24-40.6 g (13.51±0.30) respectively. The length-weight relationships were determined for females, males and all individuals as W=0.0097L2.812 (R2=0.95), W=0.0168L2.599 (R2=0.95) and W=0.0130L2.701 (R2=0.95), respectively. The growth type of A. mossulensis was negative allometric growth for female, male and all individuals. Regression analysis showed that length has high significantly correlation with weight (R = 0.97, R2 = 0.95, F1,413 = 79550.3, P <0.001) and 95% remain in weight was due to length remain. t-test results was used for the importance condition of the regression coefficients (t=89.193, P<0.01). Length-length relations were highly significant with R2 values greater than 0.97. The condition factor values were ranged between 0.430 and 0.975 for all individuals. According to the comparison of the results obtained with MAPE (%), ANNs provided better results than the LWRs. With this study, it can be seen that ANNs can be used as an alternative useful method for population parameters prediction.