Abstract

We study the changes in emotional states induced by reading and participating in online discussions, empirically testing a computational model of online emotional interaction. Using principles of dynamical systems, we quantify changes in valence and arousal through subjective reports, as recorded in three independent studies including 207 participants (110 female). In the context of online discussions, the dynamics of valence and arousal are composed of two forces: an internal relaxation towards baseline values independent of the emotional charge of the discussion, and a driving force of emotional states that depends on the content of the discussion. The dynamics of valence show the existence of positive and negative tendencies, while arousal increases when reading emotional content regardless of its polarity. The tendency of participants to take part in the discussion increases with positive arousal. When participating in an online discussion, the content of participants' expression depends on their valence, and their arousal significantly decreases afterwards as a regulation mechanism. We illustrate how these results allow the design of agent-based models to reproduce and analyze emotions in online communities. Our work empirically validates the microdynamics of a model of online collective emotions, bridging online data analysis with research in the laboratory.

Abstract

Viral marketing seeks to maximize the spread of a campaign through an online social network, often targeting influential nodes with high centrality. In this article, we analyze behavioral aspects of influential users in trust-based product reviews communities, quantifying emotional expression, helpfulness, and user activity level. We focus on two independent product review communities, Dooyoo and Epinions, in which users can write product reviews and define trust links to filter product recommendations. Following the patterns of social contagion processes, we measure user social influence by means of the k-shell decomposition of trust networks. For each of these users, we apply sentiment analysis to extract their extent of positive, negative, and neutral emotional expression. In addition, we quantify the level of feedback they received in their reviews, the length of their contributions, and their level of activity over their lifetime in the community. We find that users of both communities exhibit a large heterogeneity of social influence, and that helpfulness votes and age are significantly better predictors of the influence of an individual than sentiment. The most active of the analyzed communities shows a particular structure, in which the inner core of users is qualitatively different from its periphery in terms of a stronger positive and negative emotional expression. These results suggest that both objective and subjective aspects of reviews are relevant to the communication of subjective experience.

Abstract

Every day, millions of Internet users leave online traces that are publicly accessible. Data about forum comments, video downloads, or product reviews provide a valuable insight into human online behavior. The retrieval of datasets of unprecedented size may eventu- ally also allow testing of hypotheses or validation of theories that have been developed in the social sciences, for example, about preferences, social influence (Lorenz, 2009; Onnela & Reed-Tsochas, 2010), trust, and cooperation (Walter, Battiston, Yildirim, & Schweitzer, 2011). This recent scientific development has led to the emerging field of computational social science (Lazer et al., 2009) which combines methods and tools from different tech- nical and social disciplines. Also, psychology can benefit from this development by get- ting access to data without designing expensive experimental setups of limited size. For example, the analysis of Twitter messages allows studying the influence of the circadian cycles on human mood (Golder & Macy, 2011), and sentiment analysis of large-scale datasets reveals patterns of emotional expression predicted by earlier theories (Garcia, Garas, & Schweitzer, 2012).

Abstract

We present a study political discourse and emotional expression through a dataset of Spanish tweets. We analyze the political position of four major parties through their Twitter activity, revealing that Twitter political discourse depends on subjective perception, and resembles the political space of Spain. We propose a simplified lexicon-based method to identify the topics of a tweet, which works especially well to detect the political content of tweets. Furthermore, we adapted SentiStrength to Spanish, by translating and converting an established lexicon of word valence. Under certain design decisions, this tool performs better than random, with ample room for improvement. Finally, we combined three datasets to analyze the sentiment expressed in the political tweets of four major Spanish parties, finding differences related to the status quo, and the Spanish political climate.

Abstract

We analyse the relation between the emotions and the activity of contributors in the Open Source Software project Gentoo. Our case study builds on extensive data sets from the project's bug tracking platform Bugzilla, to quantify the activity of contributors, and its mail archives, to quantify the emotions of contributors by means of sentiment analysis. The Gentoo project is known for a considerable drop in development performance after the sudden retirement of a central contributor. We analyse how this event correlates with the negative emotions, both in bilateral email discussions with the central contributor, and at the level of the whole community of contributors. We then extend our study to consider the activity patters on Gentoo contributors in general. We find that contributors are more likely to become inactive when they express strong positive or negative emotions in the bug tracker, or when they deviate from the expected value of emotions in the mailing list. We use these insights to develop a Bayesian classifier that detects the risk of contributors leaving the project. Our analysis opens new perspectives for measuring online contributor motivation by means of sentiment analysis and for real-time predictions of contributor turnover in Open Source Software projects.

Abstract

Social organization and division of labor crucially influence the performance of collaborative software engineering efforts. In this paper, we provide a quantitative analysis of the relation between social organization and performance in Gentoo, an Open Source community developing a Linux distribution. We study the structure and dynamics of collaborations as recorded in the project's bug tracking system over a period of ten years. We identify a period of increasing centralization after which most interactions in the community were mediated by a single central contributor. In this period of maximum centralization, the central contributor unexpectedly left the project, thus posing a significant challenge for the community. We quantify how the rise, the activity as well as the subsequent sudden dropout of this central contributor affected both the social organization and the bug handling performance of the Gentoo community. We analyze social organization from the perspective of network theory and augment our quantitative findings by interviews with prominent members of the Gentoo community which shared their personal insights.

Abstract

Sentiment analysis programs are now sometimes used to detect patterns of sentiment use over time in online communication and to help automated systems interact better with users. Nevertheless, it seems that no previouspublished study has assessed whether the position of individual texts within ongoing communication can be exploited to help detect their sentiments. This article assesses apparent sentiment anomalies in on-going communication – textsassigned significantly different sentiment strength to the average of previoustexts – to see whether their classification can be improved. The results suggestthat a damping procedure to reduce sudden large changes in sentiment can improve classification accuracy but that the optimal procedure will depend on thetype of texts processed.

Abstract

We show that the frequency of word use is not only determined by the word length [1] and the average information content [2], but also by its emotional content.We have analysed three established lexica of affective word usage in English, German, and Spanish, to verify that these lexica have a neutral, unbiased, emotional content. Taking into account the frequency of word usage, we find that words with a positive emotional content are more frequently used. This lends support to Pollyanna hypothesis [3] that there should be a positive bias in human expression. We also find that negative words contain more information than positive words, as the informativeness of a word increases uniformly with its valence decrease. Our findings support earlier conjectures about (i) the relation between word frequency and information content, and (ii) the impact of positive emotions on communication and social links.

Abstract

How do users behave in online chatrooms, where they instantaneously read and write posts? We analyzed about 2.5 million posts covering various topics in Internet relay channels, and found that user activity patterns follow known power-law and stretched exponential distributions, indicating that online chat activity is not different from other forms of communication. Analysing the emotional expressions (positive, negative, neutral) of users, we revealed a remarkable persistence both for individual users and channels. I.e. despite their anonymity, users tend to follow social norms in repeated interactions in online chats, which results in a specific emotional “tone” of the channels. We provide an agent-based model of emotional interaction, which recovers qualitatively both the activity patterns in chatrooms and the emotional persistence of users and channels. While our assumptions about agent's emotional expressions are rooted in psychology, the model allows to test different hypothesis regarding their emotional impact in online communication.

Abstract

A common phenomenon on the Internet is the appearance of collective emotions, in which many users share an emotional state. Online communities allow users to emotionally interact with large amounts of other users, creating collective states faster than in offline interaction. We present our modeling framework for collective emotions in online communities. This framework allows the analysis and design of agent-based models, including the dynamics of psychological states under emotional interaction. We illustrate the applications of our framework through an overview of two different models. Based on this framework, our first model of emotions in product reviews communities reproduces the empirical distribution of emotions towards products in Amazon. The second model within our framework reproduces the emergence of emotional persistence at the individual and collective level. This persistence pattern is similar to the one revealed by our statistical analysis of IRC chatrooms. Further applications of our framework aim at reproducing collective features of emotions in a variety of online communities.

Abstract

Non-verbal communication (NVC) makes up about two-thirds of all communication between two people or between one speaker and a group of listeners. However, this fundamental aspect of communicating is mostly omitted in 3D social forums or virtual world oriented games. This paper proposes an answer by presenting a multi-user 3D-chatting system enriched with NVC relative to motion. This event-based architecture tries to recreate a context by extracting emotional cues from dialogs and derives virtual human potential body expressions from that event triggered context model. We structure the paper by expounding the system architecture enabling the modeling NVC in a multi-user 3D-chatting environment. There, we present the transition from dialog-based emotional cues to body language, and the management of NVC events in the context of a virtual reality client-server system. Finally, we illustrate the results with graphical scenes and a statistical analysis representing the increase of events due to NVC.

Abstract

This paper proposes a new emotional model for Virtual Humans (VHs) in a conversational environment. As a part of a multi-users emotional 3D-chatting system, this paper focus on how to formulate and visualize the flow of emotional state defined by the Valence-ArousalDominance (VAD) parameters. From this flow of emotion over time, we successfully visualized the change of VHs’ emotional state through the proposed emoFaces and emoMotions. The notion of Non-Verbal Communication (NVC) was exploited for driving plausible emotional expressions during conversation. With the help of a proposed interface, where a user can parameterize emotional state and flow, we succeeded to vary the emotional expressions and reactions of VHs in a 3D conversation scene.

Abstract

We analyze data about the micro-blogging site Twitter using sentiment extraction techniques. From an information per- spective, Twitter users are involved mostly in two processes: information creation and subsequent distribution (tweeting), and pure information distribution (retweeting), with pro- nounced preference to the first. However a rather substantial fraction of tweets are retweeted. Here, we address the role of the sentiment expressed in tweets for their potential after- math. We find that although the overall sentiment (polarity) does not influence the probability of a tweet to be retweeted, a new measure called emotional divergence does have an im- pact. In general, tweets with high emotional diversity have a better chance of being retweeted, hence influencing the dis- tribution of information.

Abstract

Online communities provide Internet users with means to overcome some information barriers and constraints, such as the difficulty to gather independent information about products and firms. Product review communities allow customers to share their opinions and emotions after the purchase of a product. We introduce a new dataset of product reviews from Amazon.com, with emotional information extracted by sentiment detection tools. Our statistical analysis of this data provides evidence for the existence of polemic reviews, as well as for the coexistence of positive and negative emotions inside reviews. We find a strong bias towards large values in the expression of positive emotions, while negative ones are more evenly distributed. We identified different time dynamics of the creation of reviews dependent on the presence of marketing and word of mouth effects. We define an agent-based model of the users of product review communities using a modeling framework for online emotions. This model can reproduce the scenarios of response to external influences, as well as some properties of the distributions of positive and negative emotions expressed in product reviews. This analysis and model can provide guidelines to manufacturers on how to increase customer satisfaction and how to measure the emotional impact of marketing campaigns through reviews data.

Abstract

We develop an agent-based framework to model the emergence of collective emotions, which is applied to online communities. Agent’s individual emotions are described by their valence and arousal. Using the concept of Brownian agents, these variables change according to a stochastic dynamics, which also considers the feedback from online communication. Agents generate emotional information, which is stored and distributed in a field modeling the online medium. This field affects the emotional states of agents in a non-linearmanner.We derive conditions for the emergence of collective emotions, observable in a bimodal valence distribution. Dependent on a saturated or a superlinear feedback between the information field and the agent’s arousal, we further identify scenarios where collective emotions only appear once or in a repeated manner. The analytical results are illustrated by agent-based computer simulations. Our framework provides testable hypotheses about the emergence of collective emotions, which can be verified by data from online communities.

Abstract

Assume two different communities each of which maintain their respective opinions mainly because of the weak interaction between them. In such a case, it is an interesting problem to find the necessary strength of inter-community interaction in order for the two communities to reach a consensus. In this paper, the information accumulation system (IAS) model is applied to investigate the problem. With the application of the IAS model, the opinion dynamics of the two-community problem is found to belong to a wider class of two-species problems appearing in population dynamics or in the competition of two languages, for all of which the governing equations can be described in terms of coupled logistic maps. Tipping diffusivity is defined as the maximal inter-community interaction such that the two communities maintain different opinions. For a problem with a simple community structure and homogeneous individuals, the tipping diffusivity is calculated theoretically. As a conclusion of the paper, the convergence of the two communities to the same value is less possible the more overall interaction, intra-community and inter-community, takes place. This implies, for example, that the increase in the interaction between individuals caused by the development of modern communication tools, such as Facebook and Twitter, does not necessarily improve the tendency towards global convergence between different communities. If the number of internal links increases by a factor, the number of inter-community links must be increased by an even higher factor, in order for consensus to be the only stable attractor.

Abstract

In this paper histograms of user ratings for movies (1,...,10) are analysed. The evolving stabilised shapes of histograms follow the rule that all are either double-or triple-peaked. Moreover, at most one peak can be on the central bins 2,...,9 and the distribution in these bins looks smooth `Gaussian-like’ while changes at the extremes (1 and 10) often look abrupt. It is shown that this is well approximated under the assumption that histograms are confined and discretised probability density functions of L'evy skew $α$-stable distributions. These distributions are the only stable distributions which could emerge due to a generalized central limit theorem from averaging of various independent random variables as which one can see the initial opinions of users. Averaging is also an appropriate assumption about the social process which underlies the process of continuous opinion formation. Surprisingly, not the normal distribution achieves the best fit over histograms observed on the web, but distributions with fat tails which decay as power-laws with exponent –(1 + $α$) . The scale and skewness parameters of the L'evy skew $α$-stable distributions seem to depend on the deviation from an average movie (with mean about 7.6). The histogram of such an average movie has no skewness and is the most narrow one. If a movie deviates from average the distribution gets broader and skew. The skewness pronounces the deviation. This is used to construct a one parameter fit which gives some evidence of universality in processes of continuous opinion dynamics about taste.