Analyzing and predicting sentiment of images on the social web

Siersdorfer, Stefan; Minack, Enrico; Fan, Daidu; Hare, Jonathon

doi:10.1145/1873951.1874060

Cited by 176 publications

(89 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Stefan Siersdorfer [2010] studied the association among sentiment of images expressed in metadata and their image content in the social photo sharing background Flickr. They considered the bag-of-visual words demonstration as well as the color sharing of images, and make use of the SentiWordNet lexicon to extract arithmetical values for their sentiment from related textual metadata.…”

Section: Methodsmentioning

confidence: 99%

Survey of Emotion Influence in Image Social Networks

Mettilda¹,

Anbuselvi²

2017

IJARCSSE

View full text Add to dashboard Cite

Abstract-Psychological theories propose that emotion represents the status of mind and natural responses of one's cognitive system. Emotions are a difficult state of feeling that results in physical and psychological changes that power our actions. In this paper, we study an interesting problem of emotion infection in social networks. In this paper, we study a different interesting problem of emotion influence in social networks. In particular, by employing an image social network as the basis of our study, we try to unveil how users' emotional statuses influence each other and how users' positions in the social network affect their influential strength on emotion in different papers. We also find out several interesting phenomena. For example, the possibility that a user feels happy is about linear to the number of friends who are also happy; but taking a nearer look, the pleasure chance is super linear to the number of happy friends who act as opinion leaders in the network and sub linear in the number of happy friends who span structural holes. This offers a new chance to understand the basic mechanism of emotional contagion in online social networks. Keywords-Emotion influence, image, social networks I. INTRODUCTIONWith the rapid development of online social networks such as Whatsapp, Facebook, Twitter, and Flicker, it becomes easier for people to connect with each other and share life understanding by posting messages online. In the interim, statistics show that 10% of the tweets on Twitter enclose an emotion icon. Thus an interesting question here is: will one's emotional status influence others around them? For instance, when you feel happy, will the happiness spread through your social network?Emotion influence is a development in which a person or group influences the emotions or performance of another person or group, which was previously studied by psychologists through interviews with small groups of applicants. For example, Fowler and Christakis presented the theory of three degree of authority. They find that when one feels happy, her/his friends will have a higher possibility than option to become happy. Existing research on Facebook also has established that emotion influence does happen via text-based computer-mediated communication.Comparing with text, image is a additional subjective and uncertain media for people to communicate. in addition textbased communications, will emotion influence happen in image social networks?Furthermore, researchers find that users occupying different positions in the social network participate very different roles when distribution information. For example, 1% of users acting as opinion leaders, who are taking essential positions within communities, post 50% of URLs on Twitter. In the meantime, 1% of users allocation the role of structural whole spanners, who are bridges between otherwise detached communities in a network, control 25% of information distribution. Comparing with propagating a piece of information, will these users have special patterns when propag...

show abstract

Section: Methodsmentioning

confidence: 99%

Survey of Emotion Influence in Image Social Networks

Mettilda¹,

Anbuselvi²

2017

IJARCSSE

View full text Add to dashboard Cite

show abstract

“…A standard way to address this problem leverages the supervised learning techniques with visual features extracted from a set of training data. Inspired by the success of semantic concept detection, which aims at identifying the physical appearance (e.g., object and scene) in visual instances, some handcraft low-level features (e.g., GIST and SIFT) are utilized for visual sentiment analysis in the literature [78,56,77]. A well-known problem in concept detection is semantic gap.…”

Section: Visual Sentiment Analysismentioning

confidence: 99%

Sentiment Analysis on Multi-View Social Data

Niu

Zhu

Pang

et al. 2016

Lecture Notes in Computer Science

127

View full text Add to dashboard Cite

With the proliferation of social networks, people are likely to share their opinions about news, social events and products on the Web. There is an increasing interest in understanding users' attitude or sentiment from the large repository of opinion-rich data on the Web. This can benefit many commercial and political applications. Primarily, the researchers concentrated on the documents such as users' comments on the purchased products. Recent works show that visual appearance also conveys rich human affection that can be predicted. While great efforts have been devoted on the single media, either text or image, little attempts are paid for the joint analysis of multi-view data which is becoming a prevalent form in the social media. For example, paired with the posted textual messages on Twitter, users are likely to upload images and videos which may carry their affective states. One common obstacle is the lack of sufficient manually annotated instances for model learning and performance evaluation. To prompt the researches on this problem, we introduce a multi-view sentiment analysis dataset (MVSA) including a set of manually annotated image-text pairs collected from Twitter. The dataset can be utilized as a valuable benchmark for both single-view and multi-view sentiment analysis. In this thesis, we further conduct a comprehensive study on computational analysis of sentiment from the multi-view data. The state-of-the-art approaches on single view (image or text) or multiview (image and text) data are introduced, and compared through extensive experiments conducted on our constructed dataset and other public datasets. More importantly, the effectiveness of the correlation between different views is also studied using the widely used fusion strategies and advanced multi-view feature extraction methods.Index Terms: Sentiment analysis, social media, multi-view data, textual feature, visual feature, joint feature learning.

show abstract

“…Still, there are also several recent works on sentiment analysis based on images and videos. Siersdorfer et al [14] applied machine learning techniques to predict the sentiment of images using the bag-of-visual words representation and the color distribution of images. Considering the difficulty of mapping low-level visual features to sentiment, Borth et al [1,3] and Yuan et al [15] employed attributes or entities as midlevel features to analyze visual sentiment.…”

Section: Related Workmentioning

confidence: 99%