Preliminary test to use data mining approach to manage the white clawed crayfish Austropotamobius pallipes complex in Piedmont (North-Western Italy)

In Piedmont (Italy) the impact of human beings is causing some deep environmental changes in freshwaters, to such an extent that we need to develop some practical tools to provide accurate ecological assessments of both the freshwater system and of the conditions of the species living there, one of which is the threatened crayfish Austropotamobius pallipes. We aimed at helping to manage this species by assessing its presence using two types of data-mining approaches: decision-tree models and artificial neural networks. We built models using 10 environmental input variables to classify sites as positive or negative for A. pallipes. Both decision unpruned and post-pruned tree models had high percentages of correctly classified instances and made reliable predictions. The post-pruned technique improved the reliability of the models significantly more than the unpruned, and reduced the tree complexity increasing the clarity of the models. Generally, the artificial neural network approach performed slightly worse than the decision-tree. The contribution of each independent variable in building the model was evaluated.