Balancing methods for the QT-GHT structure

The optimal balanced structure of search trees is a base requirement in efficient data management. Despite the huge work on balancing trees in vector spaces, there are very few analysis related to search trees in general metric spaces. The paper gives first an overview of the proposed quality threshold hyperplane search tree architecture in general metric space then a special method is presented for selection of pilot elements. The novelty of the proposed method is that it takes a larger set of distances into account and it provides a good solution for not uniform distributions too. The last part of the paper shows test generation and efficiency comparison.