To discover spatial association rule is one of the important contents for spatial data mining and knowledge discovery（SDMKD）

which extensively concerns with spatial

especially geographic spatial knowledge expression and reasoning.Combining with epistemic logic

formal expression is described by Inductive Logic Programming（ILP）

on the basis of geographic spatial cognition.And then the spatial predicates

possibly used in SDMKD from GIS

are analyzed and listed.Subsequently

the formal expression of spatial association rules is explained through taking examples for the relation rules between roads

rivers and towns in Suzhou region.The major problem to mine spatial association rule from GIS is to mine multi-level and multi-relational rules.And the key to solve the problem is how to effectively compute

store and express the spatial predicates among different spatial objects of different thematic layers.Firstly

the spatial data are transformed to the non-spatial data

which can be described in the related tables.And then the original problems are transformed into the problems in Boolean logic rules.Finally

on the basis of the mentioned above

according to the concept hierarchy of spatial objects

the spatial association mining approach

from top to bottom and deepening step by step

is introduced.And the spatial predicates are organized by spatial join index for predicates.