Hierarchical Clustering based on IndoorGML Document

Grouping of symbolic positions using hierarchical clustering based on IndoorGML document is presented in this paper. Using hierarchical clustering information of symbolic positions, the accuracy of symbolic indoor positioning algorithms can be improved in the case of low confidence level. The symbolic positions of the physical space can be described using IndoorGML standard. Hierarchical grouping can be determined among symbolic positions by heuristic. Although it requires knowledge gained from firsthand experience. Moreover, existing hierarchical clustering algorithms can be applied using the topological description to generate tree structures. In every tree, the leaf nodes are the symbolic positions. In this paper, these generated trees are examined.