Machine Learning Analysis of Conductivity Relative Salinity Seawater

Salinity and conductivity are fundamental parameters in understanding the physical and chemical properties of seawater, yet their complex interplay presents challenges for accurate modelling. This study investigates the relationship between salinity and conductivity in oceanographic contexts using machine learning ML techniques. The research employs a support vector machine algorithm to analyze the conductivity-salinity dataset collected from the Gulf Arabian in 2014 that focuses on exploring the effectiveness of ML models in capturing the intricate relationship between the variables and predicting salinity levels based on conductivity measurements. The result obtained a strong correlation coefficient between salinity-conductivity that hit 0.965, with a Mean Squared Error of 0.081, Mean Absolute Error of 0.172 and R-squared of 98%. While the study validation demonstrated that SVM offers robust performance than Random Forest, Decision Tree, and Linear Regression models. Our result reveals promising outcomes with ML in high prediction accuracy and effective performance in capturing nonlinear dynamic relationship that offers valuable insights for environmental monitoring, marine research, and signal propagation.