Incremental attribute computation in component-hypertrees

Abstract

Component-hypertrees are structures that store nodes of multiple component trees built with increasing neighborhoods, meaning they retain the same desirable properties of component trees but also store nodes from multiple scales, at the cost of increasing time and memory consumption for building, storing and processing the structure. In recent years, algorithmic advances resulted in optimization for both building and storing hypertrees. In this paper, we intend to further extend advances in this field, by presenting algorithms for efficient attribute computation and statistical measures that analyze how attribute values vary when nodes are merged in bigger scales. To validate the efficiency of our method, we present complexity and time consumption analyses, as well as a simple application to show the usefulness of the statistical measurements.