Uranium prediction in groundwater using support vector machine

In this article, a machine learning model (SVM) has been used to predict uranium concentration in groundwater of Gaya district. It provides a tool for uranium prediction using a set of easily measurable water quality attributes. To select the optimal inputs for the model, principal component has been used as the data reduction technique. The tuning parameter of SVM model was obtained using a grid search technique. From the results, it is evident that SVM model can be used as fast, reliable and cost-effective data analysis technique for assessment of groundwater quality. (author)