A Logical Framework for Frequent Pattern Discovery in Spatial Data

In recent times, several extensions f data mining methods and techniques have been explored aiming at dealing with advanced databases. Many promising applications of inductive logic programming (ILP) to knowledge discovery in databases have also emerged inorder to benefit from semantics andinference rules of first-order logic. Inthis paper, an ILP framework forfrequent pattern discovery in spatial data is presented. Thepattern discovery algorithm operates onfirst-order logic descriptions c mputed byan initial step of feature extraction fr m aspatial database. The algorithm benefits ofthe available background k owledge on the spatial domain and systematically explores the hierarchical structure of task-relevant geographic layers. Preliminary results have been obtained by running the algorithm SPADA onspatial data from an Italian province.