An improvement in the build algorithm for Kd-trees using mathematical mean

Querying forms a central part of dealing with data. Hence, it becomes imperative to have efficient data structures and query-search algorithms for data retrieval. Among the different types of data structures, in this paper, we have focused on 4 tree data structures. These are B, Kd, Range and Quad trees. We have made a comparative study of their build processes. Traditionally, the median of dataelements is used for the building of Kd trees. This gives a time complexity of O (k*nlogn) where n is the number of points to be sorted and k is the dimensionality of the points. We propose the building of Kd trees on the basis of mean of the data elements. The proposed algorithm has been shown to reduce the number of computations in the build process. The Kd trees produced by this methodology are balanced. Hence, this provides an improvement over the traditional building mechanism.