Intelligent Data Analysis for Diagnostics of Alcohol Dependency

The alcohol dependency is hard to diagnose since none of the existing laboratory markers has sufficient specificity and sensitivity. Therefore the goal of our study was to find better laboratory markers and / or their combinations. For that purpose the intelligent data analysis using the decision tree induction method was used. The results show that the combination of three or even two markers can prove alcohol dependency with almost 85% accuracy. However the remark has to be made that the induced decision tree offers a qualitatively different access to diagnostic evaluation of laboratory findings and varies from common practice, because it sets up its new and own borders and criteria what is the unlike from generally accepted or set up reference values. All selected markers are widely accessible, inexpensive and part of a routine laboratory tests.