Evaluation of modelling in fish length-weight relationships: The current trends

Abstract

Length and weight data for three species of fish Oreochromis niloticus, Lates niloticus, and Sydontus victorianus was collected from three East African lakes, Lake Victoria, Lake Baringo and Lake Naivasha. Correlation were done on weight and length based on the fish species, the lake collected from, sex, and the age. Correlation was found not to be a good measure of evaluating length weight relationship. Models had to be applied. Linear models, random regression models, proportional hazard models and threshold models. Random regression model was found to capture most of the information though it excluded the age of the fish on mean sum of squares error. The challenge with this is that it’s a long-term model which accounts for hatching to last date of the fish life.