<div style="text-align: justify;">In fisheries science, high number of morphometric measures (independent variables) taken from different parts of the fish complicates the estimation of the body weight (dependent variable). Therefore, the researchers are seeking for a solution facilitating the interpretation of the equations of correlation between the characteristics. One way to deal with this challenge is the dimension reduction by means of stepwise multiple regression analysis. The aim of this study is to explain total variation with the same accuracy by using fewer independent variables. To accomplish this, 12 morphometric measures from 210 individuals of Serranus cabrilla were measured to estimate the body weight. Firstly, the 95% of the variation was explained by means of multiple regression analysis by using all variables. Then, by step-wise method, the same results were achieved with fewer independent variables. Finally, the variables with inter-multicollinearity eliminated and with two remaining independent variables determination coefficients resulted as 95%. &nbsp;The result showed that using more variables does not create significant distinction for accuracy to estimate the body weight although; the total length and body dept was the most effective features for weight.&nbsp;</div>

Application of multiple regression analysis to morphometric characters obtained from <em>Serranus cabrilla </em>(linnaeus, 1758) by using stepwise method<br />&nbsp;

Application of multiple regression analysis to morphometric characters obtained from Serranus cabrilla (linnaeus, 1758) by using stepwise method