Aquaculture Water Quality Prediction using Smooth SVM

Abstract

Aquaculture, aqua farming, is the farming of aquatic organism such as fish, crustaceans, mollusk and aquatic plan. There are many factors that influence the production of aquaculture such as food stocks, protection from other predators, and water quality custody. In modern intensive river aquaculture management, water quality prediction plays an important role. The water quality indicator series are nonlinear and non-stationer. Hence, the accuracy of the commonly used conventional methods, including regression analyses and neural networks, were limited. A prediction model based on Smooth Support Vector Machine (SSVM) is proposed in this research to predict the aquaculture water quality. SSVM is an algorithm that is used for solving no linear function estimation problems. The data used in this research are data of river in Surabaya collected for two years. The data have twenty variables that indicate water quality such as temperature, turbidity, color, SS, pH, alkalinity, free CO2, DO, Nitrite, Ammonia, Copper, phosphate, sulfide, iron, Hexavalent Chromium, Manganese, Zinc, Lead, COD, and Detergents. From 520 instance data, we used 5-fold for the experiment. The Root Mean Square Error (RMSE) of the experiment is 0.0275. This value shows that SSVM proven to be an effective approach to predict aquaculture water quality.