Artificial neural networks approach in length-weight relation of crayfish (Astacus leptodactylus Eschscholtz, 1823) in Eğirdir Lake, Isparta, Turkey

Abstract

There is considerable interest in raising narrow clawed crayfish, Astacus leptodactylus (Escholtz, 1823) (A. leptodactylus), due to their high commercial value and limitations of current farming practices in dams and ponds[1]. Ecological roles of freshwater crayfish are reasonably well understood after a few decades of study. It is recognized as having an important and rather unique position in aquatic food webs[2]. It is known that freshwater habitats of crayfish are found in food webs. For this reason, they are often selected as flagship species for water protection[3]. Narrow-clawed crayfish is a widespread species distributed throughout Europe, Eastern Russia, and the Middle East[4]. The characteristics measured in these species are carapace length, body length, total length, body width, and wet weight[5]. Metric measures between individual body parts are used to show the morphological changes between gender of crayfish species[6]. If one of the metric measurements is known and the length-weighted regression can be used to calculate the length from the weight, it may be appropriate to be able to convert it to the desired length measurement[7]. Relationships between variables in these relationships are often non-linear or discovered. In regression, a transformation to achieve linearity is a special kind of nonlinear transformation. It is a nonlinear transformation that increases the linear relationship between two variables. Despite of these arrangements, the results are often inadequate and provide an insufficient forecast value for scientific studies. Besides all these results, artificial neural networks (ANNs) are emerging as nonlinear models. They do not require conversion of the parameters used and can give the desired good results[8]. Traditional statistical methods used in academic studies may be insufficient for quantification[9]. ANNs emerge as an alternative method to traditional statistical approaches for forecast modeling in nonlinear situations[10]. ANNs can be used in regression analysis involving nonlinear relations[11]. ANNs are used in prediction, classification, data association, data interpretation and data filtering processes in various disciplines of water ecology instead of biology and physical or chemical science[916]. Many studies have been carried out in forecasting studies because ARTICLE INFO ABSTRACT