Reachability vectors: features for link-based similarity measures

Abstract

In this paper, we present a novel approach to computing link-based similarities among objects accurately. We discuss the problems with previous link-based similarity measures and propose a novel approach that does not suffer from these problems. In the proposed approach, each target object is represented by a vector. The elements of the vector denote all the objects in the given data set, and the value of each element indicates the weight of the corresponding object with respect to the target object. As for this weight value, we propose to utilize the probability of reaching from the target object to the specific object, computed using the "Random Walk with Restart" strategy. Then, we define the similarity between two objects as the cosine similarity of the two vectors representing the two objects. We also evaluate the performance of the proposed approach in comparison with existing link-based measures using two kinds of data sets. Our experimental results show that the proposed approach significantly outperform the existing measures.