Abstract
This study investigates the electrical conductivity of metal halide and nitrate solutions as a function of concentration and temperature to improve electrolyte performance in electrochemical applications. Traditional electrolytes pose safety concerns due to flammability and toxicity, thus requiring alternative solutions. The research addresses the knowledge gap in understanding structure-transport relationships in ionic solutions under various conditions. Using a dataset of 100 observations, Ada Boost regression models were developed to predict conductivity behaviour. The results demonstrate strong positive correlations between concentration and conductivity for both halide (r=0.90) and nitrate (r=0.91) systems, with concentration dominating over temperature. The Ada Boost models achieved exceptional predictive accuracy (R²>0.98 for both halide and nitrate conductivity in the experimental data), with minimal error measurements confirming strong generalization. This machine learning approach successfully captures nonlinear ion transport mechanisms, providing a reliable framework for designing optimal electrolyte formulations in advanced battery and super capacitor technologies.