Figure - uploaded by Arpan Mukherjee
Content may be subject to copyright.
Performance of different regression algorithms for news-user conflict prediction.

Performance of different regression algorithms for news-user conflict prediction.

Source publication
Conference Paper
Full-text available
Over the last decade, online forums have become primary news sources for readers around the globe, and social media platforms are the space where these news forums find most of their audience and engagement. Our particular focus in this paper is to study conflict dynamics over online news articles in Reddit, one of the most popular online discussio...

Contexts in source publication

Context 1
... our dataset, the news conflict scores (computed using Eq. 3) of the news articles vary from 0 to 138.15. In Table 4, we present the MSE (Mean Squared Error), RMSE and sMAPE (Symmetric Mean Absolute Percentage Error) for predicting news conflict scores using different regression algorithms. In terms of MSE and RMSE, Lasso regression performs the best, while SVR is the best performing one when evaluated using sMAPE. ...
Context 2
... our dataset, the news conflict scores (computed using Eq. 3) of the news articles vary from 0 to 138.15. In Table 4, we present the MSE (Mean Squared Error), RMSE and sMAPE (Symmetric Mean Absolute Percentage Error) for predicting news conflict scores using different regression algorithms. In terms of MSE and RMSE, Lasso regression performs the best, while SVR is the best performing one when evaluated using sMAPE. ...

Citations

... Lai et al. [22] observed inverse homophily, where some social ties are defined as "replyto-messages" relationships. Conflicts of opinion among clustered user groups have been observed in platforms other than Twitter as well [11,21]. ...
Preprint
Full-text available
Detecting and labeling stance in social media text is strongly motivated by hate speech detection, poll prediction, engagement forecasting, and concerted propaganda detection. Today's best neural stance detectors need large volumes of training data, which is difficult to curate given the fast-changing landscape of social media text and issues on which users opine. Homophily properties over the social network provide strong signal of coarse-grained user-level stance. But semi-supervised approaches for tweet-level stance detection fail to properly leverage homophily. In light of this, We present SANDS, a new semi-supervised stance detector. SANDS starts from very few labeled tweets. It builds multiple deep feature views of tweets. It also uses a distant supervision signal from the social network to provide a surrogate loss signal to the component learners. We prepare two new tweet datasets comprising over 236,000 politically tinted tweets from two demographics (US and India) posted by over 87,000 users, their follower-followee graph, and over 8,000 tweets annotated by linguists. SANDS achieves a macro-F1 score of 0.55 (0.49) on US (India)-based datasets, outperforming 17 baselines (including variants of SANDS) substantially, particularly for minority stance labels and noisy text. Numerous ablation experiments on SANDS disentangle the dynamics of textual and network-propagated stance signals.
... Semiautomated comment moderation is one recent outcome in this field of applied research [27,29]. Modeling the dynamics of conflict in online discussions can also support moderators in deciding when to intervene [14]. ...
Conference Paper
Full-text available
The comment sections of online news platforms are an important space to indulge in political conversations andto discuss opinions. Although primarily meant as forums where readers discuss amongst each other, they can also spark a dialog with the journalists who authored the article. A small but important fraction of comments address the journalists directly, e.g., with questions, recommendations for future topics, thanks and appreciation, or article corrections. However, the sheer number of comments makes it infeasible for journalists to follow discussions around their articles in extenso. A better understanding of this data could support journalists in gaining insights into their audience and fostering engaging and respectful discussions. To this end, we present a dataset of dialogs in which journalists of The Guardian replied to reader comments and identify the reasons why. Based on this data, we formulate the novel task of recommending reader comments to journalists that are worth reading or replying to, i.e., ranking comments in such a way that the top comments are most likely to require the journalists' reaction. As a baseline, we trained a neural network model with the help of a pair-wise comment ranking task. Our experiment reveals the challenges of this task and we outline promising paths for future work. The data and our code are available for research purposes from: https://hpi.de/naumann/projects/repeatability/text-mining.html.
Article
Social networking services have been placed where people share opinions and information about various topics. These services allow users to express their opinions in direct (e.g., writing a comment or reply) and indirect ways (e.g., clicking a Like button). Based on commending, replying, and liking activities, users construct majority opinions in online environments. Previous studies examined perceptual and behavioral characteristics in the circumstance of majority opinions but only few of them provided how they differ depending on content types. Based on three different types of YouTube channels (news, documentary, and comedy), this study addresses how statistical properties of user opinions and majority opinions in online environments are presented differently depending on types of content. Based on the results of statistical analyses, we provide detailed properties of user activities in three types of YouTube channels and discuss several theoretical and practical implications.