Eleftherios Tiakas

Transitive node similarity for link prediction in social networks with positive and negative links

Online social networks (OSNs) like Facebook, and Myspace recommend new friends to registered users based on local features of the graph (i.e. based on the number of common friends that two users share). However, OSNs do not exploit the whole structure of the network. Instead, they consider only pathways of maximum length 2 between a user and his candidate friends. On the other hand, there are global approaches, which detect the overall path structure in a network, being computationally prohibitive for huge-size social networks. In this paper, we define a basic node similarity measure that captures effectively local graph features. We also exploit global graph features introducing transitive node similarity. Moreover, we derive variants of our method that apply in signed networks. We perform extensive experimental comparison of the proposed method against existing recommendation algorithms using synthetic and real data sets (Facebook, Hi5 and Epinions). Our experimental results show that our FriendTNS algorithm outperforms other approaches in terms of accuracy and it is also time efficient. We show that a significant accuracy improvement can be gained by using information about both positive and negative edges.

Searching for similar trajectories in spatial networks

In several applications, data objects move on pre-defined spatial networks such as road segments, railways, and invisible air routes. Many of these objects exhibit similarity with respect to their traversed paths, and therefore two objects can be correlated based on their motion similarity. Useful information can be retrieved from these correlations and this knowledge can be used to define similarity classes. In this paper, we study similarity search for moving object trajectories in spatial networks. The problem poses some important challenges, since it is quite different from the case where objects are allowed to move freely in any direction without motion restrictions. New similarity measures should be employed to express similarity between two trajectories that do not necessarily share any common sub-path. We define new similarity measures based on spatial and temporal characteristics of trajectories, such that the notion of similarity in space and time is well expressed, and moreover they satisfy the metric properties. In addition, we demonstrate that similarity range queries in trajectories are efficiently supported by utilizing metric-based access methods, such as M-trees.

Product recommendation and rating prediction based on multi-modal social networks

Online Social Rating Networks (SRNs) such as Epinions and Flixter, allow users to form several implicit social networks, through their daily interactions like co-commenting on the same products, or similarly co-rating products. The majority of earlier work in Rating Prediction and Recommendation of products (e.g. Collaborative Filtering) mainly takes into account ratings of users on products. However, in SRNs users can also built their explicit social network by adding each other as friends. In this paper, we propose Social-Union, a method which combines similarity matrices derived from heterogeneous (unipartite and bipartite) explicit or implicit SRNs. Moreover, we propose an effective weighting strategy of SRNs influence based on their structured density. We also generalize our model for combining multiple social networks. We perform an extensive experimental comparison of the proposed method against existing rating prediction and product recommendation algorithms, using synthetic and two real data sets (Epinions and Flixter). Our experimental results show that our Social-Union algorithm is more effective in predicting rating and recommending products in SRNs.

Trajectory Similarity Search in Spatial Networks

In several applications, data objects are assumed to move on predefined spatial networks such as road segments, railways, and invisible air routes. Moving objects may exhibit similarity with respect to their traversed paths, and therefore two objects can be correlated based on their path similarity. In this paper, we study similarity search for moving object trajectories for spatial networks. The problem poses some important challenges, since it is quite different from the case where objects are allowed to move without motion restrictions. Experimental results performed on real-life spatial networks show that trajectory similarity can be supported in an effective and efficient manner by using metric-based access methods

Transitive node similarity: predicting and recommending links in signed social networks

Online social networks (OSNs) like Facebook, Myspace, and Hi5 have become popular, because they allow users to easily share content. OSNs recommend new friends to registered users based on local features of the graph (i.e., based on the number of common friends that two users share). However, OSNs do not exploit the whole structure of the network. Instead, they consider only pathways of maximum length 2 between a user and his candidate friends. On the other hand, there are global approaches, which detect the overall path structure in a network, being computationally prohibitive for huge-size social networks. In this paper, we define a basic node similarity measure that captures effectively local graph features (i.e., by measuring proximity between nodes). We exploit global graph features (i.e., by weighting paths that connect two nodes) introducing transitive node similarity. We also derive variants of our method that apply to different types of networks (directed/undirected and signed/unsigned). We perform extensive experimental comparison of the proposed method against existing recommendation algorithms using synthetic and real data sets (Facebook, Hi5 and Epinions). Our experimental results show that our FriendTNS algorithm outperforms other approaches in terms of accuracy and it is also time efficient. Finally, we show that a significant accuracy improvement can be gained by using information about both positive and negative edges.

Progressive processing of subspace dominating queries

A top-k dominating query reports the k items with the highest domination score. Algorithms for efficient processing of this query have been recently proposed in the literature. Those methods, either index based or index free, apply a series of pruning criteria toward efficient processing. However, they are characterized by several limitations, such as (1) they lack progressiveness (they report the k best items at the end of the processing), (2) they require a multi-dimensional index or they build a grid-based index on-the-fly, which suffers from performance degradation, especially in high dimensionalities, and (3) they do not support vertically decomposed data. In this paper, we design efficient algorithms that can handle any subset of the dimensions in a progressive manner. Among the studied algorithms, the Differential Algorithm shows the best overall performance.

A unified framework for link and rating prediction in multi-modal social networks

Multi-modal social networks (MSNs) allow users to form explicit (by adding new friends in their network) or implicit (by similarly co-rating items) social networks. Previous research work was limited either to the prediction of new relationships among users (i.e., link prediction problem) or to the prediction of item ratings (i.e., rating prediction problem and item recommendations). In this paper, we develop a framework to incorporate both research directions into a unified model. Our social-union algorithm combines similarity matrices derived from heterogeneous (unipartite and bipartite) explicit or implicit MSNs. We perform an extensive experimental comparison of the proposed method against existing link and rating prediction algorithms, using synthetic and two real data sets (Epinions and Flixter). Our experimental results show that our social-union framework is more effective in both rating and link prediction.

Graph Node Clustering via Transitive Node Similarity

This paper studies the problem of cluster detection in undirected graphs by using transitive node similarity methods. Well-defined semi-metric measures are proposed to compute the similarity between the nodes of the graph, and the clustering is based on the resulted similarity values. The proposed algorithm has three major steps. In the first step, which is optional, a ranking of all the nodes of the graph is performed by using application specific criteria (if any). In the second step, a specific node is selected and the similarity values from this node to all other nodes are computed and maintained into a similarity list. In the third step, from the resulted similarity list values, the first cluster is constructed from the nodes that have a sufficient similarity score. The last two steps, are repeated, until all the nodes of the graph have been clustered. This methodology was tested in real-world data sets and provides promising clustering results. The results of a representative real-word case of clustering nodes in a real road network are presented and validated both numerically and visually.

Node and edge selectivity estimation for range queries in spatial networks

Modern applications requiring spatial network processing pose several interesting query optimization challenges. Spatial networks are usually represented as graphs, and therefore, queries involving a spatial network can be executed by using the corresponding graph representation. This means that the cost for executing a query is determined by graph properties such as the graph order and size (i.e., number of nodes and edges) and other graph parameters. In this paper, we present novel methods to estimate the number of nodes and edges in regions of interest in spatial networks, towards predicting the space and time requirements for range queries. The methods are evaluated by using real-life and synthetic data sets. Experimental results show that the number of nodes and edges can be estimated efficiently and accurately, with relatively small space requirements, thus providing useful information to the query optimizer.

Skyline queries: An introduction

During the two past decades, skyline queries were used in several multi-criteria decision support applications. Given a dominance relationship in a dataset, a skyline query returns the objects that cannot be dominated by any other objects. Skyline queries were studied extensively in multidimensional spaces, in subspaces, in metric spaces, in dynamic spaces, in streaming environments, and in time-series data. Several algorithms were proposed for skyline query processing, such as window-based, progressive, distributed, geometric-based, index-based, divide- and-conquer, and dynamic programming algorithms. Moreover, several variations were proposed to solve application-specific problems like k-dominant skylines, top-k dominating queries, spatial skyline queries, and others. As the number of objects that are returned in a skyline query may become large, there is also an extensive study for the cardinality of skyline queries. This extensive research depicts the importance of skyline queries and their variations in modern applications.