Derek O'Callaghan

An analysis of the coherence of descriptors in topic modeling

We evaluate the coherence and generality of topic descriptors found by LDA and NMF.Six new and existing corpora were specifically compiled for this evaluation.A new coherence measure using word2vec-modeled term vector similarity is proposed.NMF regularly produces more coherent topics, where term weighting is influential.NMF may be more suitable for topic modeling of niche or non-mainstream corpora. In recent years, topic modeling has become an established method in the analysis of text corpora, with probabilistic techniques such as latent Dirichlet allocation (LDA) commonly employed for this purpose. However, it might be argued that adequate attention is often not paid to the issue of topic coherence, the semantic interpretability of the top terms usually used to describe discovered topics. Nevertheless, a number of studies have proposed measures for analyzing such coherence, where these have been largely focused on topics found by LDA, with matrix decomposition techniques such as Non-negative Matrix Factorization (NMF) being somewhat overlooked in comparison. This motivates the current work, where we compare and analyze topics found by popular variants of both NMF and LDA in multiple corpora in terms of both their coherence and associated generality, using a combination of existing and new measures, including one based on distributional semantics. Two out of three coherence measures find NMF to regularly produce more coherent topics, with higher levels of generality and redundancy observed with the LDA topic descriptors. In all cases, we observe that the associated term weighting strategy plays a major role. The results observed with NMF suggest that this may be a more suitable topic modeling method when analyzing certain corpora, such as those associated with niche or non-mainstream domains.

Online social media in the Syria conflict: encompassing the extremes and the in-betweens

The Syria conflict has been described as the most socially mediated in history, with online social media playing a particularly important role. At the same time, the ever-changing landscape of the conflict leads to difficulties in applying analytical approaches taken by other studies of online political activism. Therefore, in this paper, we use an approach that does not require strong prior assumptions or the proposal of an advance hypothesis to analyze Twitter and YouTube activity of a range of protagonists to the conflict, in an attempt to reveal additional insights into the relationships between them. By means of a network representation that combines multiple data views, we uncover communities of accounts falling into four categories that broadly reflect the situation on the ground in Syria. A detailed analysis of selected communities within the anti-regime categories is provided, focusing on their central actors, preferred online platforms, and activity surrounding “real world” events. Our findings indicate that social media activity in Syria is considerably more convoluted than reported in many other studies of online political activism, suggesting that alternative analytical approaches can play an important role in this type of scenario.

Uncovering the wider structure of extreme right communities spanning popular online networks

Recent years have seen increased interest in the online presence of extreme right groups. Although originally composed of dedicated websites, the online extreme right milieu now spans multiple networks, including popular social media platforms such as Twitter, Facebook and YouTube. Ideally therefore, any contemporary analysis of online extreme right activity requires the consideration of multiple data sources, rather than being restricted to a single platform. We investigate the potential for Twitter to act as one possible gateway to communities within the wider online network of the extreme right, given its facility for the dissemination of content. A strategy for representing heterogeneous network data with a single homogeneous network for the purpose of community detection is presented, where these inherently dynamic communities are tracked over time. We use this strategy to discover and analyze persistent English and German language extreme right communities.

Down the White Rabbit Hole

In addition to hosting user-generated video content, YouTube provides recommendation services, where sets of related and recommended videos are presented to users, based on factors such as co-visitation count and prior viewing history. This article is specifically concerned with extreme right (ER) video content, portions of which contravene hate laws and are thus illegal in certain countries, which are recommended by YouTube to some users. We develop a categorization of this content based on various schema found in a selection of academic literature on the ER, which is then used to demonstrate the political articulations of YouTube’s recommender system, particularly the narrowing of the range of content to which users are exposed and the potential impacts of this. For this purpose, we use two data sets of English and German language ER YouTube channels, along with channels suggested by YouTube’s related video service. A process is observable whereby users accessing an ER YouTube video are likely to be recommended further ER content, leading to immersion in an ideological bubble in just a few short clicks. The evidence presented in this article supports a shift of the almost exclusive focus on users as content creators and protagonists in extremist cyberspaces to also consider online platform providers as important actors in these same spaces.

Characterizing ego-networks using motifs

We assess the potential of network motif profiles to characterize ego-networks in much the same way that a bag-of-words strategy allows text documents to be compared in a vector space framework. This is potentially valuable as a generic strategy for comparing nodes in a network in terms of the network structure in which they are embedded. In this paper, we consider the computational challenges and model selection decisions involved in network motif profiling. We also present three case studies concerning the analysis of Wikipedia edit networks, YouTube spam campaigns, and peer-to-peer lending in the Prosper marketplace.