The length-weight relationship of fish is considered to be a critical tool in describing the key biological aspects of fish stocks: estimating the fish weight on the basis of length measurements and vice versa; conducting growth pattern analysis using the allometric coefficient of the analyzed species; carrying out analysis of body condition of the sampled fish specimens. Moreover, the knowledge (and data) acquired as a result of the study into the functional relationship between the length and the weight of fish is of paramount importance for the provision of fish stock assessments, for developing fish stock condition analysis and performing strict fisheries monitoring programs.
For the purposes of the present research the length-weight relationship is investigated through the use of a MATLAB based specific software developed to perform the analysis in two stages: the first one involves direct application of linear regression analysis on data sets of length-weight measurements of sprat and anchovy commercial catch samples and the second one validates the non-linear length-weight relationship model W(i) = q *L(i)b for both species. The present analysis shows that both linear and non-linear models are valid and statistically significant for the analyzed species under certain conditions. Presented are relevant conclusions, direct comparison of accuracy and analysis of the results obtained by applying the models described above.