Published on January 2021 | Aquaculture fisheries, Regression modeling

Relative Importance of Water Quality Parameters on Fish Growth Using PLS Regression

Partial least squares (PLS) is a multivariate dimension reduction technique which is not based on ordinary regression assumptions. The use of PLS regression in life sciences is still a novel concept despite having many scientific applications. This paper analyses the relative importance of physicochemical parameters on the growth of Oreochromis jipe using Oreochromis niloticus as a control. Modelling and the graphical display of the regression coefficients were performed using a suite of open access R-software packages. The modelling hypotheses were assessed using experimental data collected from 270 fingerlings cultured for the period of 84 days. The findings revealed that significant linear correlation exists between water quality variables and the mean body weight of both O. jipe and O. niloticus fish species. The study provides baseline information to assess the growth of O. jipe under aquaculture conditions; therefore, we recommend a further study to be conducted on several other predictor variables that can be measured under controlled aquaculture conditions.