Sentiment Expression Research Articles

Modeling Arabic subjectivity and sentiment in lexical space

In spite of the vast amount of work on subjectivity and sentiment analysis (SSA), it is not yet particularly clear how lexical information can best be modeled in a morphologically-richness language. To bridge this gap, we report successful models targeting lexical input in Arabic, a language of very complex morphology. Namely, we measure the impact of both gold and automatic segmentation on the task and build effective models achieving significantly higher than our baselines. Our models exploiting predicted segments improve subjectivity classification by 6.02% F1-measure and sentiment classification by 4.50% F1-measure against the majority class baseline surface word forms. We also perform in-depth (error) analyses of the behavior of the models and provide detailed explanations of subjectivity and sentiment expression in Arabic against the morphological richness background in which the work is situated.

Sentiment Analysis of Movie Review using Machine Learning Approach

With development of Internet and Natural Language processing, use of regional languages is also grown for communication. Sentiment analysis is natural language processing task that extracts useful information from various data forms such as reviews and categorize them on basis of polarity. One of the sub-domain of opinion mining is sentiment analysis which is basically focused on the extraction of emotions and opinions of the people towards a particular topic from textual data. In this paper, sentiment analysis is performed on IMDB movie review database. We examine the sentiment expression to classify the polarity of the movie review on a scale of negative to positive and perform feature extraction and ranking and use these features to train our multilevel classifier to classify the movie review into its correct label. In this paper classification of movie reviews into positive and negative classes with the help of machine learning. Proposed approach using classification techniques has the best accuracy of about 99%.

The Limits of Material Benefits: Remittances and Pro-Americanism in Mexico

We explore how the reception of remittances affects perceptions of the bilateral relationship between Mexico and the United States. Scholars have claimed that the economic benefits of the relationship with the US prevail over imperialistic concerns as a result of the asymmetry of power between the two countries. Empirical research shows that Latin American public opinion is indeed more supportive of the US than theory indicates. However, we identify two gaps in this literature. First, scholars have explored the determinants of generic expressions of sentiment toward the US, overlooking more concrete instances of cooperation between the two countries. Second, scholars have focused on trade and investment and have ignored how the material gains of emigration shape attitudes toward the US. The present paper fills these two gaps by using novel survey data on the bilateral relationship between Mexico and the US. On one hand, we find that while the reception of remittances correlates positively with good sentiments toward the US, the recipients of remittances are consistently more opposed to cooperation with the US in the fight against drug trafficking. We argue that this finding can be explained by the different nature of the migratory phenomenon, and the connection between anti-drug trafficking policies and the close scrutiny of illegal flows of money and people.

Knowledge management in consultancy involved LPS implementation projects via social media

Lean production systems (LPS) have been adopted by numerous organizations to improve their performance. Many enterprise implement LPS by inviting external consultants and using social media to promote this collaboration. Social media provide a channel for enterprise employee and external consultants to communicate and exchange their respective knowledge as a form of their reaction to others or surrounding (Hines et al. in Am Antiq 23(3):312, 29; Hu et al. in Prod Plan Control 27:550–562, 31; Kipping and Armbruster in The burden of otherness: limits of consultancy interventions in historical case studies, Buchkapitel, Routledge, 37). Using a combination of content analysis and sentiment analysis, this study comprehensively analyzed the effect of sentiment expression to knowledge interaction taking place on social media. Three major influences could be culled from the paper. The case study implies that, through social media, both consultant and organizational employee could play a more proactive role in the knowledge interaction, and sentiment expression makes it more effective.

Multi-modality weakly labeled sentiment learning based on Explicit Emotion Signal for Chinese microblog

Understanding the sentiments of users from cross media contents which contain texts and images is an important task for many social network applications. However, due to the semantic gap between cross media features and sentiments, machine learning methods need a lot of human labeled samples. Furthermore, for each kind of media content, it is necessary to constantly add a lot of new human labeled samples because of new expressions of sentiments. Fortunately, there are some emotion signals, like emoticons, which denote users’ emotions in cross media contents. In order to use these weakly labels to build a unified multi-modality sentiment learning framework, we propose an Explicit Emotion Signal (EES) based multi-modality sentiment learning approach which uses huge number of weakly labeled samples in sentiment learning. There are three advantages in our approach. Firstly, only a few human labeled samples are needed to reach the same performance which can be obtained by the traditional machine learning based sentiment prediction approaches. Secondly, this approach is flexible and can easily combine text and vision based sentiment learning through deep neural networks. Thirdly, because a lot of weakly labeled samples can be used in EES, trained model is more robust in different domain transfer. In this paper, firstly, we investigate the correlation between sentiments and emoticons and choose emoticons as the Explicit Emotion Signals in our approach; secondly, we build a two stages multi-modality sentiment learning framework based on Explicit Emotion Signals. Our experiment results show that our approach not only achieves the best performance but also only needs 3% and 43% training samples to obtain the same performance of Visual Geometry Group (VGG) model and Long Short-Term Memory (LSTM) model in images and texts, respectively.

How do Indonesians remember Konfrontasi? Indonesia–Malaysia relations and the popular memory of “Confrontation” after the fall of Suharto

ABSTRACTAfter the fall of Suharto in May 1998, mass rallies yelling anti-Malaysia sentiment broke out several times in a number of major cities in Indonesia. The rallies were triggered by various conflicting issues involving the two countries. Every time a mass rally against Malaysia happens, memory of “Konfrontasi” is recalled, as is seen in the use of “Ganyang Malaysia” (Crush Malaysia) rhetoric, whereas during the Suharto era, the narrative of the historical episode of “Konfrontasi” was constructed in the tone of criticizing Sukarno’s “Crush Malaysia” campaign as an escape from the internal economic crisis, rather than as an expression of nationalist sentiment. However, as this article addresses, there is a gap between the “national memory” as is constructed by the history school textbook and “popular memory” as is embodied in society. Beneath this “popular memory,” as this paper contends, there is a sort of nationalist sentiment in the sense of longing for “national pride” as projected upon the “persona” of Sukarno.

Collaboratively Training Sentiment Classifiers for Multiple Domains

We propose a collaborative multi-domain sentiment classification approach to train sentiment classifiers for multiple domains simultaneously. In our approach, the sentiment information in different domains is shared to train more accurate and robust sentiment classifiers for each domain when labeled data is scarce. Specifically, we decompose the sentiment classifier of each domain into two components, a global one and a domain-specific one. The global model can capture the general sentiment knowledge and is shared by various domains. The domain-specific model can capture the specific sentiment expressions in each domain. In addition, we extract domain-specific sentiment knowledge from both labeled and unlabeled samples in each domain and use it to enhance the learning of domain-specific sentiment classifiers. Besides, we incorporate the similarities between domains into our approach as regularization over the domain-specific sentiment classifiers to encourage the sharing of sentiment information between similar domains. Two kinds of domain similarity measures are explored, one based on textual content and the other one based on sentiment expressions. Moreover, we introduce two efficient algorithms to solve the model of our approach. Experimental results on benchmark datasets show that our approach can effectively improve the performance of multi-domain sentiment classification and significantly outperform baseline methods.

Personality-based refinement for sentiment classification in microblog

Microblog has become one of the most widely used social media for people to share information and express opinions. As information propagates fast in social network, understanding and analyzing public sentiment implied in user-generated content is beneficial for many fields and has been applied to applications such as social management, business and public security. Most previous work on sentiment analysis makes no distinctions of the tweets by different users and ignores the diverse word use of people. As some sentiment expressions are used by specific groups of people, the corresponding textual sentiment features are often neglected in the analysis process. On the other hand, previous psychological findings have shown that personality influences the ways people write and talk, suggesting that people with same personality traits tend to choose similar sentiment expressions. Inspired by this, in this paper we propose a method to facilitate sentiment classification in microblog based on personality traits. To this end, we first develop a rule-based method to predict users’ personality traits based on the most well-studied personality model, the Big Five model. In order to leverage more effective but not widely used sentiment features, we then extract those features grouped by different personality traits and construct personality-based sentiment classifiers. Moreover, we adopt an ensemble learning strategy to integrate traditional textual feature based and our personality-based sentiment classification. Experimental studies on Chinese microblog dataset show the effectiveness of our method in refining the performance of both the traditional and state-of-the-art sentiment classifiers. Our work is among the first to explicitly explore the role of user's personality in social media analytics and its application in sentiment classification.

User-aware topic modeling of online reviews

The online reviews are one type of social media which are opinions generated by the users to comment on some special items. Since the sentiments are dependent on topics, probabilistic topic models have been widely used for sentiment analysis. However, most of existing methods only model the text, but rarely consider the users, who express the opinions, and the items, which the opinions are expressed on. Different users are usually concerned with different topics and use different sentiment expressions, a lenient user might tend to give positive review than a critical user. High-quality items tend to receive positive reviews than low-quality items. To better model the topics and sentiments, we argue that it is essential to explore reviews as well as users and items. To this end, we propose a novel model called User Item Sentiment Topic (UIST) which incorporates users and items for topic modeling and produces topic–word, user–topic, and item–topic distributions simultaneously. Extensive experiments on several datasets demonstrate the advantages and effectiveness of our method. The extracted topics with our method are more coherent and informative; consequently, the performance of sentiment classification is also improved. Furthermore, the user preference obtained with our method could be utilized for many personalized applications.

“My Profession Chose Me:” Precarity, Pragmatism, and Professionalism in Caribbean and West African Immigrant Domestic Workers’ Narratives

AbstractExtensive literature on care work or domestic work — childcare, eldercare, and housecleaning — in the United States and internationally has highlighted how the sector's affective aspects, e.g., intimacy, personalism, and fictive kin relationships, extract “surplus emotional labor” (Hochschild 1983) from women workers, who, in turn, often justify their caregiving labor through expressions of sentiment and kin obligation. In this article, I draw on three women's narrative framings of their domestic work jobs to argue that immigrant women in New York City emphasize structural, practical, and contextual elements of their domestic/caregiving labor, challenging constructions of care sector labor as evidence of innate caring aptitudes, womanly duty, or moral service. These characterizations have direct labor policy implications for paid care work — as evidenced in recent reforms. They also provide a theoretical basis for broader redistributive politics linking paid and unpaid caregiving labor.

Understanding Health Care Social Media Use From Different Stakeholder Perspectives: A Content Analysis of an Online Health Community.

BackgroundHealth care social media used for health information exchange and emotional communication involves different types of users, including patients, caregivers, and health professionals. However, it is difficult to identify different stakeholders because user identification data are lacking due to privacy protection and proprietary interests. Therefore, identifying the concerns of different stakeholders and how they use health care social media when confronted with huge amounts of health-related messages posted by users is a critical problem.ObjectiveWe aimed to develop a new content analysis method using text mining techniques applied in health care social media to (1) identify different health care stakeholders, (2) determine hot topics of concern, and (3) measure sentiment expression by different stakeholders.MethodsWe collected 138,161 messages posted by 39,606 members in lung cancer, diabetes, and breast cancer forums in the online community MedHelp.org over 10 years (January 2007 to October 2016) as experimental data. We used text mining techniques to process text data to identify different stakeholders and determine health-related hot topics, and then analyzed sentiment expression.ResultsWe identified 3 significantly different stakeholder groups using expectation maximization clustering (3 performance metrics: Rand=0.802, Jaccard=0.393, Fowlkes-Mallows=0.537; P<.001), in which patients (24,429/39,606, 61.68%) and caregivers (12,232/39,606, 30.88%) represented the majority of the population, in contrast to specialists (2945/39,606, 7.43%). We identified 5 significantly different health-related topics: symptom, examination, drug, procedure, and complication (Rand=0.783, Jaccard=0.369, Fowlkes-Mallows=0.495; P<.001). Patients were concerned most about symptom topics related to lung cancer (536/1657, 32.34%), drug topics related to diabetes (1883/5904, 31.89%), and examination topics related to breast cancer (8728/23,934, 36.47%). By comparison, caregivers were more concerned about drug topics related to lung cancer (300/2721, 11.03% vs 109/1657, 6.58%), procedure topics related to breast cancer (3952/13,954, 28.32% vs 5822/23,934, 24.33%), and complication topics (4449/25,701, 17.31% vs 4070/31,495, 12.92%). In addition, patients (9040/36,081, 25.05%) were more likely than caregivers (2659/18,470, 14.39%) and specialists (17,943/83,610, 21.46%) to express their emotions. However, patients’ sentiment intensity score (2.46) was lower than those of caregivers (4.66) and specialists (5.14). In particular, for caregivers, negative sentiment scores were higher than positive scores (2.56 vs 2.18), with the opposite among specialists (2.62 vs 2.46). Overall, the proportion of negative messages was greater than that of positive messages related to symptom, complication, and examination. The pattern was opposite for drug and procedure topics. A trend analysis showed that patients and caregivers gradually changed their emotional state in a positive direction.ConclusionsThe hot topics of interest and sentiment expression differed significantly among different stakeholders in different disease forums. These findings could help improve social media services to facilitate diverse stakeholder engagement for health information sharing and social interaction more effectively.

Entropy based classifier for cross-domain opinion mining

In recent years, the growth of social network has increased the interest of people in analyzing reviews and opinions for products before they buy them. Consequently, this has given rise to the domain adaptation as a prominent area of research in sentiment analysis. A classifier trained from one domain often gives poor results on data from another domain. Expression of sentiment is different in every domain. The labeling cost of each domain separately is very high as well as time consuming. Therefore, this study has proposed an approach that extracts and classifies opinion words from one domain called source domain and predicts opinion words of another domain called target domain using a semi-supervised approach, which combines modified maximum entropy and bipartite graph clustering. A comparison of opinion classification on reviews on four different product domains is presented. The results demonstrate that the proposed method performs relatively well in comparison to the other methods. Comparison of SentiWordNet of domain-specific and domain-independent words reveals that on an average 72.6% and 88.4% words, respectively, are correctly classified.

Exchange Cards: Advertising, Album Making, and the Commodification of Sentiment in the Gilded Age

This article examines the materiality of advertising trade cards used in the Gilded Age United States and resituates the medium within the post–Civil War culture of sentimental and personal exchange. Mobilizing evidence from over 3,000 cards and numerous scrapbooks, this article demonstrates that market culture commodified sentimental images and themes in chromolithographed cards, enabling consumers to appropriate them for sentimental expression in albums. The function and usage patterns of such cards as album keepsakes thus illustrates an underlying tension in the nineteenth-century ideals of separate spheres and sincere expression.

Unveiling What Is Written in the Stars: Analyzing Explicit, Implicit, and Discourse Patterns of Sentiment in Social Media

Abstract Deciphering consumers’ sentiment expressions from big data (e.g., online reviews) has become a managerial priority to monitor product and service evaluations. However, sentiment analysis, the process of automatically distilling sentiment from text, provides little insight regarding the language granularities beyond the use of positive and negative words. Drawing on speech act theory, this study provides a fine-grained analysis of the implicit and explicit language used by consumers to express sentiment in text. An empirical text-mining study using more than 45,000 consumer reviews demonstrates the differential impacts of activation levels (e.g., tentative language), implicit sentiment expressions (e.g., commissive language), and discourse patterns (e.g., incoherence) on overall consumer sentiment (i.e., star ratings). In two follow-up studies, we demonstrate that these speech act features also influence the readers’ behavior and are generalizable to other social media contexts, such as Twitter and Facebook. We contribute to research on consumer sentiment analysis by offering a more nuanced understanding of consumer sentiments and their implications.

The False Idol of Beauty: Bl. John Henry Newman's Critiques of Aesthetics and the Challenge of Evangelization

The False Idol of BeautyBl. John Henry Newman's Critiques of Aesthetics and the Challenge of Evangelization Matthew Muller (bio) Charles Taylor has shown how, in the romantic period, there arose a need to discover and express, on an individualistic basis, one's authentic human life. For the romantics, "nature as source" meant finding meaning and identity within the individual, and the most authentic experience of nature was the experience of sentiment. As Taylor notes, the romantics gave "a central and positive place to sentiments in the moral life," such that they took "the centrality of sentiment to unheard-of lengths."1 The need to construct an identity from the inner source of nature required attentiveness to one's sentiments or emotions. "Fulfilling my nature means espousing the inner élan, the voice or impulse," and this "expressive view of human life went along naturally with a new understanding of art."2 Throughout the late eighteenth and into the nineteenth century, continuing still today, argues Taylor, is a dissolution of "the distinction between the ethical and the aesthetic."3 The definition of aesthetics likewise changed in this time, so that, "along with a new understanding of natural and artistic beauty," aesthetics "focused less on the nature of the object and more on the quality of the experience evoked."4 If we accept Taylor's narrative and description of the modern person's [End Page 54] understanding of beauty, art, and aesthetics, then there are serious challenges to the possibility that beauty can serve as a useful source of evangelization. If the experience of beauty is merely about my experience of sentiment, and the creation of the beautiful is about the manifestation of individualistic and subjective sentiments, then it need not draw a person towards wonder, questioning, or search. In this article, I argue that Bl. John Henry Newman came to view aesthetics in the way that Taylor describes, but that he also recognized the pitfalls of art as sentiment in cultivating Christian faith, virtue, and devotion. I argue that Newman understood artistic production, most often spoken of in relation to poetry, as the expression of sentiment, or the release of sentiment. Since the goal of artistic production was the communication of sentiment, the experience of sentiment was likewise the primary purpose and effect of viewing art. This led Newman to be very cautious about the experience of beauty as a means of evangelization, or, to speak more in his terms, as a means for growth in virtue and religious duty and devotion. To show this, I would like to give a brief account of what we might call Newman's aesthetic theory. Of particular importance is Newman's assertion that beauty is self-sufficient, in that it does not point beyond itself, which is consistent with Taylor's account of expressivism. This then, as I will argue, has a formative impact on the imagination, so that depending on the perspective one brings to his or her experience of beauty, the imagination may be prejudiced against Christian faith. In this sense, taste or aesthetics becomes the standard for truth or goodness. In other words, something might be considered good or evil, true or false, based on the criteria of emotional experience caused by the perception of beauty in the object, action, or event. Newman's concern for the effects of beauty on an individual's religious commitments still hold today if art, in its production, but more importantly, in its reception (the way it is viewed) is about merely the experience of sentiments. The challenge is increased today, when a whole range of experiences of beauty are available through print and social media. [End Page 55] The first point to note is that Newman and his fellow Tractarians received from the romantic movement the conception of art as the expression of personal sentiment. One of the main sources of Newman's aesthetic theory was John Keble, whom Newman called "the true and primary author" of the Oxford Movement.5 Keble published a volume of poetry, The Christian Year, in 1827, which gave the Oxford Movement its spiritual ethos.6 Over a decade earlier, however, Keble had already articulated what...

Muslims at the Australian periphery

The overt expression of anti-Muslim sentiment is a relatively new phenomenon in Australia. It builds upon racism embedded in history, “clash of civilisations” ideologies and constructs of border-terrorism. Denigration of Muslims, commonly termed Islamophobia, is overtly evident in the official sphere, media reporting and increasing popular rejection of Islamic amenities such as schools and mosques. Connected but more insidious is the Islamophobia of the ‘white savior rescue’ movement, in which Muslim men and Islam are positioned as perpetrators of oppression and harm toward Muslim women, requiring non-Muslim intervention. Varied forms of Islamophobia and their impacts are discussed.

Россияне и жители Ростовской области о настроениях и повседневных практиках в условиях кризиса 2016 года: соотношение общероссийского и регионального

The article brings into focus the social sentiments of the population defining the content and direction of changes occurring in society, as well as analyzes adaptive practices that allow residents of the region to integrate into a dynamic social reality and participate in the development of macrosocial processes. The article highlights the expression of social sentiments among residents of Rostov Region in times of crisis in 2016, which are characterized by a decrease in the level of optimism, awareness of the long-term crisis and the growth of unresolved problems within society. While assessing the prospects for development of society, the dominant trend was revealed — the growth of emotional tension and anxiety in public expectations. Based on analysis of empirical data, the strategies for adapting the residents of the Rostov Region to the social and economic crisis were considered. In particular, it was found that residents of the region are characterized by a fairly low level of variability in the willingness to implement adaptation practices. A distinctive feature of regional respondents is revealed: relying on one’s own resources, the independent search for options for adaptation to crisis aspects, and the refusal to seek help from not only state, but also social and religious organizations. The survey was carried out according to methods developed by the scientists of the IS RAS, and as for a number of thematic units, the results are given in comparison with all-Russian data.

An unsupervised machine learning model for discovering latent infectious diseases using social media data

IntroductionThe authors of this work propose an unsupervised machine learning model that has the ability to identify real-world latent infectious diseases by mining social media data. In this study, a latent infectious disease is defined as a communicable disease that has not yet been formalized by national public health institutes and explicitly communicated to the general public. Most existing approaches to modeling infectious-disease-related knowledge discovery through social media networks are top-down approaches that are based on already known information, such as the names of diseases and their symptoms. In existing top-down approaches, necessary but unknown information, such as disease names and symptoms, is mostly unidentified in social media data until national public health institutes have formalized that disease. Most of the formalizing processes for latent infectious diseases are time consuming. Therefore, this study presents a bottom-up approach for latent infectious disease discovery in a given location without prior information, such as disease names and related symptoms. MethodsSocial media messages with user and temporal information are extracted during the data preprocessing stage. An unsupervised sentiment analysis model is then presented. Users’ expressions about symptoms, body parts, and pain locations are also identified from social media data. Then, symptom weighting vectors for each individual and time period are created, based on their sentiment and social media expressions. Finally, latent-infectious-disease-related information is retrieved from individuals’ symptom weighting vectors. Datasets and resultsTwitter data from August 2012 to May 2013 are used to validate this study. Real electronic medical records for 104 individuals, who were diagnosed with influenza in the same period, are used to serve as ground truth validation. The results are promising, with the highest precision, recall, and F1 score values of 0.773, 0.680, and 0.724, respectively. ConclusionThis work uses individuals’ social media messages to identify latent infectious diseases, without prior information, quicker than when the disease(s) is formalized by national public health institutes. In particular, the unsupervised machine learning model using user, textual, and temporal information in social media data, along with sentiment analysis, identifies latent infectious diseases in a given location.

Affective Sovereignty in Louis XIII's Royal Ballet

In early modern France, there was a critical tension between monarchical conduct, the development of Absolutism and a concurrent emergence of what the anglophone world would today call ‘emotion’. This article would like to explore this tension from the perspective of performance studies by looking at royal ballet as a form of affective sovereignty. Although royal ballet provided the perfect conditions for kings and their court members to stage sovereignty, under Louis XIII there is an unusual disruption of this view of the ballet as the public transmission of an ideal crown. In an alarming break from ideals of royal dignity, Louis and his fellow dancers surprise the court with their attachment to dance and their affection (and repulsion) for each other, leaving the ballet librettist scrambling to spin the ballet programnes into politically viable expressions of sovereign sentiment.

Targeted Twitter Sentiment Analysis for Brands Using Supervised Feature Engineering and the Dynamic Architecture for Artificial Neural Networks

Social media communications offer valuable feedback to firms about their brands. We present a targeted approach to Twitter sentiment analysis for brands using supervised feature engineering and the dynamic architecture for artificial neural networks. The proposed approach addresses challenges associated with the unique characteristics of the Twitter language and brand-related tweet sentiment class distribution. We demonstrate its effectiveness on Twitter data sets related to two distinctive brands. The supervised feature engineering for brands offers final tweet feature representations of only seven dimensions with greater feature density. Reducing the dimensionality of the representations reduces the complexity of the classification problem and feature sparsity. Two sets of experiments are conducted for each brand in three-class and five-class tweet sentiment classification. We examine five-class classification to target the mild sentiment expressions that are of particular interest to firms and brand management practitioners. We compare the proposed approach to the performances of two state-of-the-art Twitter sentiment analysis systems from the academic and commercial domains. The results indicate that it outperforms these state-of-the-art systems by wide margins, with classification F1-measures as high as 88 percent and excellent recall of tweets expressing mild sentiments. Furthermore, they demonstrate the tweet feature representations, though consisting of only seven dimensions, are highly effective in capturing indicators of Twitter sentiment expression. The proposed approach and vast majority of features identified through supervised feature engineering are applicable across brands, allowing researchers and brand management practitioners to quickly generate highly effective tweet feature representations for Twitter sentiment analysis on other brands.