ABSTRACT
In the past a few decades, the decision tree (DT) induction method is one of the most prevalent and powerful techniques in data mining. As for the defects of traditional DTs, Zhou et al. presented a co-location-based decision tree (CL-DT) method to enhance the decision making, and this has been successfully applied in pavement maintenance strategies. The major characteristics of CL-DT consider the geospatial relationship of these attribute data in addition to the traditional attribute data and use a co-location mining technology to first classify the co-location attributes. The basic idea of a CL-DT is to apply co-location rules to induce the generation of a DT. The processes of CL-DT consist of (a) selecting nonspatial and spatial data, (b) determining rough candidate co-locations, (c) determining candidate co-locations, (d) pruning the non-prevalent co-locations, (e) inducing co-location rules, (f) formulating a node merging criterion, and (g) inducing a co-location decision tree.
