Link analysis ranking: algorithms, theory, and experiments

Abstract

The explosive growth and the widespread accessibility of the Web has led to a surge of research activity in the area of information retrieval on the World Wide Web. The seminal papers of Kleinberg [1998, 1999] and Brin and Page [1998] introduced Link Analysis Ranking, where hyperlink structures are used to determine the relative authority of a Web page and produce improved algorithms for the ranking of Web search results. In this article we work within the hubs and authorities framework defined by Kleinberg and we propose new families of algorithms. Two of the algorithms we propose use a Bayesian approach, as opposed to the usual algebraic and graph theoretic approaches. We also introduce a theoretical framework for the study of Link Analysis Ranking algorithms. The framework allows for the definition of specific properties of Link Analysis Ranking algorithms, as well as for comparing different algorithms. We study the properties of the algorithms that we define, and we provide an axiomatic characterization of the INDEGREE heuristic which ranks each node according to the number of incoming links. We conclude the article with an extensive experimental evaluation. We study the quality of the algorithms, and we examine how different structures in the graphs affect their performance.

What’s in a Cluster? Automatically Detecting Interesting Interactions in Student E-Discussions

“…We compared the performance of DOCE using different feature sets of the algorithm and also compared it to a random algorithm, as there is no other comparable "gold standard" algorithm, at least not for the particular type of problem we are tackling in this work. Our methodology is similar to Information Retrieval (IR) evaluations -the "Top10" results are considered in the evaluation and the relevancy of results is defined based on user feedback [18]. As already explained, DOCE does not always match clusters in an exact manner.…”

Efficient Ranking on Entity Graphs with Personalized Relationships

“…In ObjectRank, P specifies the nodes that contain the query keywords (all nodes for Global ObjectRank). In addition to ObjectRank, which is an extension of the PageRank authority flow ranking method, HITS [19] or its extensions [20] may also be used for authority flow-based personalization.…”

Graph-based methods for the automatic annotation and retrieval of art prints

“…Graph-based learning has been thoroughly studied by the information retrieval community to rank web pages on the World Wide Web [6,8,35]. Essentially, a graph is built where vertexes represent web pages and the edge weights are denoted by the existence of hyper-links.…”

“…Note that this assumption is important to uncover the aforementioned influence of art prints over other artistic productions (and also over other art prints). Specifically, we explore the following graph-based algorithms [6,35,50]: label propagation [18,28,34,46], random walk [11], stationary solution using a stochastic matrix [30], and combinatorial harmonic [19]. We adapt each of those techniques to a bag of visual words (BOV) approach, and we compare their performance with BOV approaches that use the following classifiers: support vector machines (SVM) [45], and random forests (RF) [7].…”