Voronoi Diagram Based Retrieval Method for the Internet of Things

Abstract

In the Internet of Things, search service is that information which is corresponding to entities in real world is retrieve so as to help people achieve information that they need. From the angle of visualization, this paper study the search service in the Internet of Things. With the Vector-Space Model, entities' information is expressed as entity vectors, and a Voronoi diagram based Vector-Space Model visualization method is presented. According to similarity between entity vectors, entity vectors' position in two-dimensional plane is calculated and this plane is divided into Voronoi diagram by nodes as which feature vectors of entity vectors’ clustering are taken. According to similarity between query vector and feature vector, a Voronoi diagram based Vector-Space Model retrieval method is put forward. This method restricts the searching scope in Voronoi domain of feature vectors that are the most similar to query vector, thereby the number of compared entity vectors is reduced in the retrieval process. The experiment result indicates that this method can ensure retrieval precision and improve retrieval efficiency.