Abstract | ||
---|---|---|
In this paper, we describe two different learning tasks for relational structures. When learning a classifier for st ructures, the relational structures in the training sets are classifie d as a whole. Contrarily, when learning a context dependent classifier fo r elemen- tary objects, the elementary objects of the relational stru ctures in the training set are classified. We investigate the question how such classifications can be induced automatically from a given tr aining set containing classified structures or classified elementary o bjects re- spectively. We present an algorithm based on fast graph isomorphism testing that allows the description of the objects in the tra ining set by automatically constructed attributes. This allows us to employ well- known methods of decision tree induction to construct a hypothesis. We describe new simplification and structure reconstructio n tech- niques for the learned structural decision tree. We present the system INDIGO and evaluate it on the Mesh and the Mutagenicity Data. |
Year | Venue | Keywords |
---|---|---|
1996 | ECAI | relational learning,graph isomorphism,context dependent,decision tree |
Field | DocType | Citations |
Information Fuzzy Networks,Decision tree,Statistical relational learning,Computer science,Influence diagram,Artificial intelligence,ID3 algorithm,Alternating decision tree,Machine learning,Decision tree learning,Incremental decision tree | Conference | 7 |
PageRank | References | Authors |
0.69 | 7 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Peter Geibel | 1 | 286 | 26.62 |
Fritz Wysotzki | 2 | 456 | 45.46 |