IMT Institutional Repository: No conditions. Results ordered -Date Deposited. 2024-03-28T09:39:08ZEPrintshttp://eprints.imtlucca.it/images/logowhite.pnghttp://eprints.imtlucca.it/2014-12-03T11:12:58Z2014-12-18T13:55:51Zhttp://eprints.imtlucca.it/id/eprint/2387This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/23872014-12-03T11:12:58ZConsistent community identification in complex networksWe have found that known community identification algorithmsproduce inconsistent communities when the node ordering changes atinput. We propose two metrics to quantify the level of consistencyacross multiple runs of an algorithm: pairwise membershipprobability and consistency. Based on these two metrics, weaddress the consistency problem without compromising themodularity. Our solution uses pairwise membership probabilitiesas link weights and generates consistent communities within six orfewer cycles. It offers a new tool in the study of communitystructures and their evolutions.Haewoon KwakSue MoonYoung-Ho Eomyoungho.eom@imtlucca.itYoonchan ChoiHawoong Jeong2014-12-02T15:15:35Z2014-12-18T13:56:23Zhttp://eprints.imtlucca.it/id/eprint/2384This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/23842014-12-02T15:15:35ZMining communities in networksOnline social networks pose significant challenges to computer scientists, physicists, and sociologists alike, for their massive size, fast evolution, and uncharted potential for social computing. One particular problem that has interested us is community identification. Many algorithms based on various metrics have been proposed for communities in networks [18, 24], but a few algorithms scale to very large networks. Three recent community identification algorithms, namely CNM [16], Wakita [59], and Louvain [10], stand out for their scalability to a few millions of nodes. All of them use modularity as the metric of optimization. However, all three algorithms produce inconsistent communities every time the ordering of nodes to the algorithms changes.
We propose two quantitative metrics to represent the level of consistency across multiple runs of an algorithm: pairwise membership probability and consistency. Based on these two metrics, we propose a solution that improves the consistency without compromising the modularity. We demonstrate that our solution to use pairwise membership probabilities as link weights generates consistent communities within six or fewer cycles for most networks. However, our iterative, pairwise membership reinforcing approach does not deliver convergence for Flickr, Orkut, and Cyworld networks as well for the rest of the networks. Our approach is empirically driven and is yet to be shown to produce consistent output analytically. We leave further investigation into the topological structure and its impact on the consistency as future work.
In order to evaluate the quality of clustering, we have looked at 3 of the 48 communities identified in the AS graph. Surprisingly, all have either hierarchical, geographical, or topological interpretations to their groupings. Our preliminary evaluation of the quality of communities is promising. We plan to conduct more thorough evaluation of the communities and study network structures and their evolutions using our approach.
Haewoon KwakYoonchan ChoiYoung-Ho Eomyoungho.eom@imtlucca.itHawoong JeongSue Moon2014-12-02T15:12:14Z2014-12-18T13:56:58Zhttp://eprints.imtlucca.it/id/eprint/2383This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/23832014-12-02T15:12:14ZComparison of online social relations in volume vs interaction
Online social networking services are among the most popular Internet services according to Alexa.com and have become a key feature in many Internet services. Users interact through various features of online social networking services: making friend relationships, sharing their photos, and writing comments. These friend relationships are expected to become a key to many other features in web services, such as recommendation engines, security measures, online search, and personalization issues. However, we have very limited knowledge on how much interaction actually takes place over friend relationships declared online. A friend relationship only marks the beginning of online interaction.
Does the interaction between users follow the declaration of friend relationship? Does a user interact evenly or lopsidedly with friends? We venture to answer these questions in this work. We construct a network from comments written in guestbooks. A node represents a user and a directed edge a comments from a user to another. We call this network an activity network. Previous work on activity networks include phone-call networks [34, 35] and MSN messenger networks [27]. To our best knowledge, this is the first attempt to compare the explicit friend relationship network and implicit activity network.
We have analyzed structural characteristics of the activity network and compared them with the friends network. Though the activity network is weighted and directed, its structure is similar to the friend relationship network. We report that the in-degree and out-degree distributions are close to each other and the social interaction through the guestbook is highly reciprocated. When we consider only those links in the activity network that are reciprocated, the degree correlation distribution exhibits much more pronounced assortativity than the friends network and places it close to known social networks. The k-core analysis gives yet another corroborating evidence that the friends network deviates from the known social network and has an unusually large number of highly connected cores.
We have delved into the weighted and directed nature of the activity network, and investigated the reciprocity, disparity, and network motifs. We also have observed that peer pressure to stay active online stops building up beyond a certain number of friends.
The activity network has shown topological characteristics similar to the friends network, but thanks to its directed and weighted nature, it has allowed us more in-depth analysis of user interaction.
Hyunwoo ChunHaewoon KwakYoung-Ho Eomyoungho.eom@imtlucca.itYong-Yeol AhnSue MoonHawoong Jeong