Multi-label Classification of pH Levels using Support Vector Machines

This paper developed an intelligent system application for the multi-label classification of pH levels. The pH is a measure of how acidic or how basic a substance is. The use of supervised learning methods may serve as a cheaper and more reliable alternative for pH level measurement. In this study, hue-saturation-value (HSV) color data were used for the training and testing the developed model. The obtained dataset has four field attributes including the output. Support vector machine (SVM) classification was the supervised learning tool used to model the classification system. 1410 samples from the dataset were used for the training (987 samples) and the testing (423 samples). Moreover, several kernel functions such as polynomial and radial basis function (RBF) kernel were examined when designing the classification system. Model evaluation through metric functions show that the trained SVM with a polynomial kernel has a 99.41% accuracy. As a result, the developed model was able to produce multiple decision hyperplanes for the multi-label classification task.