Using truss networks to estimate the biomass of Oreochromis niloticus, and to investigate shape Characteristics

Truss and length measurements were made on 463 tilapia Oreochromis niloticus of known mass. These measurements were then used to build mathematical models to estimate tilapia biomass. Two different approaches were used to build the models. The first was to take some simple mathematical models, and to optimize them using least squares regression. The second approach was to build models using artificial neural network software. Models were built for two different but overlapping size ranges and both approaches were successful in creating models that estimated the total biomass to within ±1%. The truss measurements were also used to assess the principal shape components of different sized tilapia.