Predicting chemical parameters of river water quality from bioindicator data

We address the problem of inferring chemical parameters of river water qual ity from biological ones. This task is important for enabling selective che mical monitoring of river water quality. We apply machine learning, in part icular regression tree induction, to biological and chemical data on the wa ter quality of Slovenian rivers. Regression trees are constructed that pred ict values of chemical parameters from data on the presence of bioindicator taxa at the species and family levels.