Citation Context

...group might be considered (i) as a set of nodes (i.e., users) more densely connected among each other than to the others (i.e., the group formation is viewed as a graph clustering problem [13], [14], =-=[20]-=-); or, (ii) as a community of P. De Meo is with the Dept. of Ancient and Modern Civilizations, University of Messina, 98166 Messina, Italy, e-mail: pdemeo@unime.it E. Ferrara is with the Center for Co...

Citation Context

...users [27]. However, if we assume that homogeneity should reflect users satisfaction, we argue that other behavioral characteristics of members and groups should be considered as important components =-=[8]-=-. For example, in virtual communities, users are often characterized by multiple interests, and groups enact common rules, define accepted behaviors, exhibit a manifold of communication styles and imp...

Citation Context

...nowledge, no study considers the evolution of a group as a problem of matching between users and groups profiles. Although the concept of social profile is known in the context of virtual communities =-=[25]-=-, that of group profile is rather novel. The definition of such concept is useful to face the problem of suggesting a user the groups she could affiliate to, so that to improve her satisfaction. Commo...

Citation Context

... domenico.rosaci@unirc.it G.M.L. Sarné is with the Dept. DICEAM, University of Reggio Calabria, Loc. Feo di Vito, 89122 Reggio Calabria, Italy, e-mail: sarne@unirc.it people sharing similar interests =-=[12]-=- (i.e., the group formation accounts for some definition of users similarity). Satisfaction, on the other hand, is often related to the notion of group homogeneity: when the similarity/inter-connectiv...

Citation Context

...n recommendation in [39]). This differs from the group recommendation problem where the objects to recommend are items whereas the affiliation recommendation problem deals with groups. Spertus et al. =-=[38]-=- presented a proposal that describes an empirical comparison of six distinct measures for computing the similarity of a user and a community to exploit for communities recommendation. Chen et al. [9] ...

Citation Context

...tent [5], [18], [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], =-=[21]-=-, [28], [29], [37]. Recent studies investigated the relationships between users and groups in OSNs [2], [23], [24]. For example, Hui et al. [23] considered four popular OSNs and empirically computed t...

Citation Context

...ty/inter-connectivity among group participants is high, according to both structural and semantic dimensions, a OSN group is regards as homogeneous and this yields better satisfaction among its users =-=[27]-=-. However, if we assume that homogeneity should reflect users satisfaction, we argue that other behavioral characteristics of members and groups should be considered as important components [8]. For e...

Citation Context

...d by this paper, illustrating the main novelties brought in by our approach. In the latest years, an increasing number of authors focused on the problem of recommending items to the member of a group =-=[1]-=-, [22]. This implies the need to construct a group profile, often by simply aggregating the individual orientations of its members. This task is usually called group modelling. More formally, let U, I...

Citation Context

... to the average of the ratings the member of G have given to ik. If none of the users in G has rated in ik, then fG(ik) is set equal to ⊥ (this symbol specifying a not rated item)); (ii) Least Misery =-=[3]-=-, where the rating that group G would assign to ik is defined as fG(ik) = min r(ui, ik) if ∃ui ∈ U : r(ui, ik) = ⊥ and ⊥ otherwise. In the Average strategy the score of an item ik depends on how many...

Citation Context

...Group Homogeneity. I. INTRODUCTION Online Social Networks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], [7], significantly grown in scale and content [5], [18], =-=[26]-=-, with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [3...

Citation Context

...[38] presented a proposal that describes an empirical comparison of six distinct measures for computing the similarity of a user and a community to exploit for communities recommendation. Chen et al. =-=[9]-=- provide an algorithm called CCF (Combinational Collaborative Filtering) which is able to suggest users new friendship relationships as well as the communities they could join with. CCF considers a co...

Citation Context

...-agent systems, Online Social Networks, Group Recommendation, Group Homogeneity. I. INTRODUCTION Online Social Networks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities =-=[6]-=-, [7], significantly grown in scale and content [5], [18], [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-netwo...

Citation Context

...tion, Group Homogeneity. I. INTRODUCTION Online Social Networks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], [7], significantly grown in scale and content [5], =-=[18]-=-, [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [2...

Citation Context

...cial Networks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], [7], significantly grown in scale and content [5], [18], [26], with significant social effects [10], =-=[11]-=-, [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [37]. Recent studies investigated the relation...

Citation Context

.... This definition of group profile is coherent with the definition of a user profile containing information comparable with those of a group profile. • Each OSN group is associated with a group agent =-=[15]-=-– [17], capable of creating, managing and updating the group profile defined above. Similarly, a user agent is associated with each OSN user. • We present a distributed agent platform to handle group ...

Citation Context

...aling with the matching of a user and a group profile. Most of this work has been designed to recommend to an OSN user groups to join with (such a problem is also called affiliation recommendation in =-=[39]-=-). This differs from the group recommendation problem where the objects to recommend are items whereas the affiliation recommendation problem deals with groups. Spertus et al. [38] presented a proposa...

Citation Context

...this paper, illustrating the main novelties brought in by our approach. In the latest years, an increasing number of authors focused on the problem of recommending items to the member of a group [1], =-=[22]-=-. This implies the need to construct a group profile, often by simply aggregating the individual orientations of its members. This task is usually called group modelling. More formally, let U, I and G...

Citation Context

...vs epochs obtained with the GHMcomp and GHM-diff algorithms, for a SN with 30.000 users and 100 groups. To measure the internal homogeneity of a group g we use the average homogeneity AHg, derived by =-=[33]-=-, computed as ∑ x,y∈g,x=y hx,y/|g|, while to measure the global homogeneity of the OSN groups we compute the mean average homogeneity MAH and the standard deviation average homogeneity DAH of all the...

Citation Context

...s (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], [7], significantly grown in scale and content [5], [18], [26], with significant social effects [10], [11], [19], =-=[30]-=-. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [37]. Recent studies investigated the relationships betwee...

Citation Context

...mendation, Group Homogeneity. I. INTRODUCTION Online Social Networks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], [7], significantly grown in scale and content =-=[5]-=-, [18], [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [2...

Citation Context

...ine Social Networks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], [7], significantly grown in scale and content [5], [18], [26], with significant social effects =-=[10]-=-, [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [37]. Recent studies investigated the re...

Citation Context

...etworks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], [7], significantly grown in scale and content [5], [18], [26], with significant social effects [10], [11], =-=[19]-=-, [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [37]. Recent studies investigated the relationships ...

Citation Context

...le is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [37]. Recent studies investigated the relationships between users and groups in OSNs [2], =-=[23]-=-, [24]. For example, Hui et al. [23] considered four popular OSNs and empirically computed the probability that a user joins a group; the problem of choosing which group to join has been studied in [2...

Citation Context

...t systems, Online Social Networks, Group Recommendation, Group Homogeneity. I. INTRODUCTION Online Social Networks (OSNs) such as Facebook, Google+ and Twitter have become very complex realities [6], =-=[7]-=-, significantly grown in scale and content [5], [18], [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks o...

Citation Context

...played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [37]. Recent studies investigated the relationships between users and groups in OSNs [2], [23], =-=[24]-=-. For example, Hui et al. [23] considered four popular OSNs and empirically computed the probability that a user joins a group; the problem of choosing which group to join has been studied in [2] for ...

Citation Context

...nt role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], [37]. Recent studies investigated the relationships between users and groups in OSNs =-=[2]-=-, [23], [24]. For example, Hui et al. [23] considered four popular OSNs and empirically computed the probability that a user joins a group; the problem of choosing which group to join has been studied...

Citation Context

...Commonly, a group might be considered (i) as a set of nodes (i.e., users) more densely connected among each other than to the others (i.e., the group formation is viewed as a graph clustering problem =-=[13]-=-, [14], [20]); or, (ii) as a community of P. De Meo is with the Dept. of Ancient and Modern Civilizations, University of Messina, 98166 Messina, Italy, e-mail: pdemeo@unime.it E. Ferrara is with the C...

Citation Context

...ly, a group might be considered (i) as a set of nodes (i.e., users) more densely connected among each other than to the others (i.e., the group formation is viewed as a graph clustering problem [13], =-=[14]-=-, [20]); or, (ii) as a community of P. De Meo is with the Dept. of Ancient and Modern Civilizations, University of Messina, 98166 Messina, Italy, e-mail: pdemeo@unime.it E. Ferrara is with the Center ...

Citation Context

...d content [5], [18], [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests =-=[4]-=-, [21], [28], [29], [37]. Recent studies investigated the relationships between users and groups in OSNs [2], [23], [24]. For example, Hui et al. [23] considered four popular OSNs and empirically comp...

Citation Context

...of creating, managing and updating the group profile defined above. Similarly, a user agent is associated with each OSN user. • We present a distributed agent platform to handle group formation [31], =-=[32]-=-, [34]–[36]. The agents automatically and dynamically compute a matching between user and group profiles in a distributed fashion. We provide the user agent with a matching algorithm, named Group Homo...

Citation Context

..., managing and updating the group profile defined above. Similarly, a user agent is associated with each OSN user. • We present a distributed agent platform to handle group formation [31], [32], [34]–=-=[36]-=-. The agents automatically and dynamically compute a matching between user and group profiles in a distributed fashion. We provide the user agent with a matching algorithm, named Group Homogeneity Max...

Citation Context

...5], [18], [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], =-=[28]-=-, [29], [37]. Recent studies investigated the relationships between users and groups in OSNs [2], [23], [24]. For example, Hui et al. [23] considered four popular OSNs and empirically computed the pro...

Citation Context

...8], [26], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], =-=[29]-=-, [37]. Recent studies investigated the relationships between users and groups in OSNs [2], [23], [24]. For example, Hui et al. [23] considered four popular OSNs and empirically computed the probabili...

Citation Context

...ating, managing and updating the group profile defined above. Similarly, a user agent is associated with each OSN user. • We present a distributed agent platform to handle group formation [31], [32], =-=[34]-=-–[36]. The agents automatically and dynamically compute a matching between user and group profiles in a distributed fashion. We provide the user agent with a matching algorithm, named Group Homogeneit...

Citation Context

...6], with significant social effects [10], [11], [19], [30]. In this context, a relevant role is played by social groups, that are sub-networks of users sharing common interests [4], [21], [28], [29], =-=[37]-=-. Recent studies investigated the relationships between users and groups in OSNs [2], [23], [24]. For example, Hui et al. [23] considered four popular OSNs and empirically computed the probability tha...

Citation Context

... definition of group profile is coherent with the definition of a user profile containing information comparable with those of a group profile. • Each OSN group is associated with a group agent [15]– =-=[17]-=-, capable of creating, managing and updating the group profile defined above. Similarly, a user agent is associated with each OSN user. • We present a distributed agent platform to handle group format...

Citation Context

...pable of creating, managing and updating the group profile defined above. Similarly, a user agent is associated with each OSN user. • We present a distributed agent platform to handle group formation =-=[31]-=-, [32], [34]–[36]. The agents automatically and dynamically compute a matching between user and group profiles in a distributed fashion. We provide the user agent with a matching algorithm, named Grou...