This paper presents a new map representing the structure of all of science, based on journal articles, including both the natural and social sciences. Similar to cartographic maps of our world, the map of science provides a bird's eye view of today's scientific landscape. It can be used to visually identify major areas of science, their size, similarity, and interconnectedness. In order to be useful, the map needs to be accurate on a local and on a global scale. While our recent work has focused on the former aspect, this paper summarizes results on how to achieve structural accuracy. Eight alternative measures of journal similarity were applied to a data set of 7,121 journals covering over 1 million documents in the combined Science Citation and Social Science Citation Indexes. For each journal similarity measure we generated two-dimensional spatial layouts using the force-directed graph layout tool, VxOrd. Next, mutual information values were calculated for each graph at different clustering levels to give a measure of structural accuracy for each map. The best co-citation and inter-citation maps according to local and structural accuracy were selected and are presented and characterized. These two maps are compared to establish robustness. The inter-citation map is more » then used to examine linkages between disciplines. Biochemistry appears as the most interdisciplinary discipline in science. « less

Mapping the backbone of science.

International Joint Conference on Neural Networks

https://cyberleninka.org/article/n/854933.pdf

Social media analytics – Challenges in topic discovery, data collection, and data preparation

We survey the several measures that exists in literature to rank influential users in Twitter network.We propose a classification of these measures according to different criteria, such as the kind of metrics, the use of PageRank algorithm, the use of content analysis, among others.Besides the influential users, we also study measures of activity and popularity.We finish by mentioning some aspects of this topic related with computational complexity and correlation measures. Centrality is one of the most studied concepts in social network analysis. There is a huge literature regarding centrality measures, as ways to identify the most relevant users in a social network. The challenge is to find measures that can be computed efficiently, and that can be able to classify the users according to relevance criteria as close as possible to reality. We address this problem in the context of the Twitter network, an online social networking service with millions of users and an impressive flow of messages that are published and spread daily by interactions between users. Twitter has different types of users, but the greatest utility lies in finding the most influential ones. The purpose of this article is to collect and classify the different Twitter influence measures that exist so far in literature. These measures are very diverse. Some are based on simple metrics provided by the Twitter API, while others are based on complex mathematical models. Several measures are based on the PageRank algorithm, traditionally used to rank the websites on the Internet. Some others consider the timeline of publication, others the content of the messages, some are focused on specific topics, and others try to make predictions. We consider all these aspects, and some additional ones. Furthermore, we include measures of activity and popularity, the traditional mechanisms to correlate measures, and some important aspects of computational complexity for this particular context.

Measuring user influence on Twitter

Human language can be described as a complex network of linked words. In such a treatment, each distinct word in language is a vertex of this web, and neighboring words in sentences are connected by edges. It was recently found (Ferrer and Sol\'e) that the distribution of the numbers of connections of words in such a network is of a peculiar form which includes two pronounced power-law regions. Here we treat language as a self-organizing network of interacting words. In the framework of this concept, we completely describe the observed Word Web structure without fitting.

Language as an Evolving Word Web

In this paper we consider the community detection problem from two different perspectives. We first want to be able to compute communities for large directed networks, containing million vertices and billion arcs. Moreover, in a large number of applications, the graphs modelizing such networks are directed. Nevertheless, one is often forced to forget the direction between the connections, either for the sake of simplicity or because no other options are available. This is in particular the case on large networks, since there are only a few scalable algorithms at the time. We thus turn our attention to one of the most famous scalable algorithms, namely Louvain's algorithm [3], based on modularity maximization. We modify Louvain's algorithm to handle directed networks based on the notion of directed modularity defined by Leicht and Newman [13], and provide an empirical and theoretical study to show that one should prefer directed modularity. To illustrate this fact, we use the LFR benchmarks by Lancichinetti and Fortunato [8] to design an evaluation benchmark of directed graphs with community structure. We also give some examples and insights on the situations where one should really consider direction when maximizing modularity. Finally, for the sake of completeness, we compare the results obtained with Oslom [12], one of the best algorithms to detect communities in directed networks. While the results obtained with such an algorithm are by far better on the LFR benchmarks, we emphasize that it is still not well-suited to deal with very large networks.

https://hal.archives-ouvertes.fr/hal-01231784/document

Directed Louvain : maximizing modularity in directed networks

In this paper, we investigate the issue of detecting the real-life influence of people based on their Twitter account. We propose an overview of common Twitter features used to characterize such accounts and their activity, and show that these are inefficient in this context. In particular, retweets and followers numbers, and Klout score are not relevant to our analysis. We thus propose several Machine Learning approaches based on Natural Language Processing and Social Network Analysis to label Twitter users as Influencers or not. We also rank them according to a predicted influence level. Our proposals are evaluated over the CLEF RepLab 2014 dataset, and outmatch state-of-the-art ranking methods.

Detecting Real-World Influence through Twitter

In this paper, we focus on the detection and behavior of social capitalists, a special kind of users in Twitter. Roughly speaking, social capitalists follow users regardless of their contents, just hoping to increase their number of followers. They have first been introduced by Ghosh et al. (Proceedings of the 21st international conference on World Wide Web, WWW’12, pp. 61–70, 2012). In this work, we provide a method to detect these users efficiently. Our algorithms do not rely on the tweets posted by the users, just on the topology of the Twitter graph. Then, we show that these users form a highly connected group in the network by studying their neighborhoods and their local clustering coefficient (Watts and Strogatz, Nature 393 (6684):440–442, 1998). We next study the evolution of such users between 2009 and 2013. Finally, we provide a behavioral analysis based on social capitalists that tweet on a special hashtag. Our work emphasizes that such users, who act like automatic accounts, are in fact for most of them real users.

Social capitalists on Twitter: detection, evolution and behavioral analysis

Many works related to Twitter aim at characterizing its users in some way: role on the service (spammers, bots, organizations, etc.), nature of the user (socio-professional category, age, etc.), topics of interest, and others. However, for a given user classification problem, it is very difficult to select a set of appropriate features, because the many features described in the literature are very heterogeneous, with name overlaps and collisions, and numerous very close variants. In this article, we review a wide range of such features. In order to present a clear state-of-the-art description, we unify their names, definitions and relationships, and we propose a new, neutral, typology. We then illustrate the interest of our review by applying a selection of these features to the offline influence detection problem. This task consists in identifying users who are influential in real life, based on their Twitter account and related data. We show that most features deemed efficient to predict online influence, such as the numbers of retweets and followers, are not relevant to this problem. However, we propose several content-based approaches to label Twitter users as influencers or not. We also rank them according to a predicted influence level. Our proposals are evaluated over the CLEF RepLab 2014 dataset, and outmatch state-of-the-art methods.

A review of features for the discrimination of twitter users: application to the prediction of offline influence

This paper deals with a major challenge in clustering that is optimal model selection. It presents new efficient clustering quality indexes relying on feature maximization, which is an alternative measure to usual distributional measures relying on entropy, Chi-square metric or vector-based measures such as Euclidean distance or correlation distance. First Experiments compare the behavior of these new indexes with usual cluster quality indexes based on Euclidean distance on different kinds of test datasets for which ground truth is available. This comparison clearly highlights altogether the superior accuracy and stability of the new method on these datasets, its efficiency from low to high dimensional range and its tolerance to noise. Further experiments are then conducted on “real life” textual data extracted from a multisource bibliographic database for which ground truth is unknown. These experiments show that the accuracy and stability of these new indexes allow to deal efficiently with diachronic analysis, when other indexes do not fit the requirements for this task.

/pdf/new-efficient-clustering-quality-indexes-3wazccff9q.pdf

Nicolas Dugué

Papers

Directed Louvain : maximizing modularity in directed networks

Detecting Real-World Influence through Twitter

Social capitalists on Twitter: detection, evolution and behavioral analysis

A review of features for the discrimination of twitter users: application to the prediction of offline influence

New efficient clustering quality indexes