The Researches on Knowledge Distance and Its Relative Extensions

Abstract

All knowledge in the universe constructs a metric space, and the properties of knowledge distance are studied. Moreover, the knowledge distance was used to measure the rough entropy of knowledge. The concept of relative knowledge distance is based on the principle of relative cognition to embody the difficulty of transforming cognition between two knowledge under reference knowledge. It proved that the relative knowledge distance is monotonous with the refinement of reference knowledge granularity. In the experiment, the structural characteristics of relative knowledge distance in hierarchical clustering are visually analyzed. The feature selection algorithm based on relative knowledge distance is designed to simulate the process of human conditional cognition, which clarified that due to different cognitive perspectives and cognitive paths, relative knowledge distance has shown different cognitive characteristics such as enhancement, maintenance, and weakening in problem-solving.