No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
User Sentiment Detection: A YouTube Use Case
... There are many web platforms that are used to share non-textual content such as videos, images, animations that allow users to add comments for each item [1]. YouTube is probably the most popular of them, with millions of videos uploaded by its users and billions of comments for all of these videos. ...
... The specific list contains terms and phrases. These terms and phrases express user opinions [1,13].The SentiWordNet (SWN) is a document resource which contains a list of English terms which have been attributed a score of positivity and negativity [14].The SWN assigns polarity to each term and phrase of WordNet [12,15].For analyzing user sentiment in comments, SentiWordNet [6,9,16 ] is used. Through sentiment analysis positive and negative opinions, emotions can be identified. ...
... Through sentiment analysis positive and negative opinions, emotions can be identified. User comments can be analyzed by using SWN and customized social media specific phrase list [1,17]. ...
Sentiment analysis or opinion mining is the field of study related to analyze
opinions, sentiments, evaluations, attitudes, and emotions of users which they
express on social media and other online resources. The revolution of social
media sites has also attracted the users towards video sharing sites, such as
YouTube. The online users express their opinions or sentiments on the videos
that they watch on such sites. This paper presents a brief survey of techniques
to analyze opinions posted by users about a particular video.
... Keyword-based Features [15] Continuous-bag of words [18] Pattern, Skip-gram [15,18] Ngram, Word2Vec [18][19][20] Sentiment, Lexicon, Polarity detriment to the individual who is abused, but it can also have a negative impact on community relations. Abusive posts have become a major problem for social media platforms like Twitter. ...
... Reference Approach [15,[18][19][20] Keyword-based Approaches [2,10,22,11,27,28] Content-based Approaches [3,31,4,8,30,38] Context-based Approaches [5,43,[45][46][47][48]54,55] Other Approaches A probabilistic clustering method along with fuzzy classification is used to detect hate speech [45] on Twitter. Other approaches use deep learning and pre-trained embeddings to identify and visualize hate on Twitter [46] and Facebook [46,47]. ...
This paper presents a novel approach for detecting abuse on Twitter. Abusive posts have become a major problem for social media platforms like Twitter. It is important to identify abuse to mitigate its potential harm. Many researchers have proposed methods to detect abuse on Twitter. However, most of the existing approaches for detecting abuse look only at the content of the abusive tweet in isolation and do not consider its contextual information, particularly the tweets posted before the abusive tweet. In this paper, we propose a new method for detecting abuse that uses contextual information from the tweets that precede and follow the abusive tweet. We hypothesize that this contextual information can be used to better understand the intent of the abusive tweet and to identify abuse that content-based methods would otherwise miss. We performed extensive experiments to identify the best combination of features and machine learning algorithms to detect abuse on Twitter. We test eight different machine learning classifiers on content- and context-based features for the experiments. The proposed method is compared with existing abuse detection methods and achieves an absolute improvement of around 7%. The best results are obtained by combining the content and context-based features. The highest accuracy of the proposed method is 86%, whereas the existing methods used for comparison have highest accuracy of 79.2%.
... Context is not limited to specific parameters it can be represented by different parameters which vary based upon the condition of the recommender systems. [15] The major purpose of using contextual information in recommendation is to convert two dimensional (2D) recommender systems into three dimensional (3D) recommender systems as shown in Eq. (2). Traditional 2D collaborative-filtering based recommender systems are represented through Eq. (1) [15], ...
... [15] The major purpose of using contextual information in recommendation is to convert two dimensional (2D) recommender systems into three dimensional (3D) recommender systems as shown in Eq. (2). Traditional 2D collaborative-filtering based recommender systems are represented through Eq. (1) [15], ...
YouTube is one of the most popular video sharing website being used by the users throughout the world. For providing ease to the user it offers a list of recommended videos every time the user searches some content. But many times the provided or recommended videos are not related to context that the user had searched. This is due to the title and the description of the videos which are although related to the keyword that the user had searched but the content of the video may be different. Moreover, the videos are recommended on the premise of users' interest irrespective of the context they are in. Therefore, the recommended videos cover different interests of the user altogether. The existing approaches are predominantly based on content and collaborative recommendations. So in this research work, the proposed and recommended approach is context based. The recommended videos are to be positioned on the basis of association and comment feedback. Moreover, for improving the quality of ranking, structural analysis (i.e. Meta information about the videos) is also performed on each video to get high relativity videos.
... The context represents the physical or mental state that the user is currently in. Different parameters can be used to represent the context of a user or the item that is being accessed [11]. The basic purpose of the context-aware recommender systems is to convert a 2D recommender system into a 3D recommender system. ...
... The basic purpose of the context-aware recommender systems is to convert a 2D recommender system into a 3D recommender system. Collaborative recommender systems can be represented as [11], ...
... RELATED [7] shows that with hyperparameter optimization distributional models performance can be improved and also experiment design choice can be an important factor. Smitashree Choudhury, John G. Breslin [8] created data corpus using YouTube API and used Porter stemming approach and SentiWordnet for the task detecting the polarity of the content. Alok Gupta, Dipeeka Sonavane, Kajal Attarde, Neelam Shelar, Pramila Mate [9] used a technique for polarity detection of movie reviews in Hindi and achieved a significant performance. ...
Cyberbullying has become a major problem nowadays especially on social networking sites. Training a system to detect the negative comments to remove them from the sites can solve this problem. Analyzing the attitude of the comments or posts on social media using word embeddings can identify it. word embeddings with a simple dimensionality reduction operation called Hellinger PCA (Principal Component Analysis) is used in this paper. The dataset is created by collecting comments from popular pages of Facebook using Graph API. On detecting cyberbullying and negative comments, accuracy of 72% is obtained for our corpus of 10500 with this approach. The performance is improving for larger corpus.
... For precise sentiment inference from comments, it is essential to integrate the video content. Current approaches tend to treat comment sentiment analysis as a simple NLP task and neglect the semantic connection between videos and comments [48,44,33,10,2,27]. ...
Existing video multi-modal sentiment analysis mainly focuses on the sentiment expression of people within the video, yet often neglects the induced sentiment of viewers while watching the videos. Induced sentiment of viewers is essential for inferring the public response to videos and has broad application in analyzing public societal sentiment, effectiveness of advertising and other areas. The micro videos and the related comments provide a rich application scenario for viewers' induced sentiment analysis. In light of this, we introduces a novel research task, Multi-modal Sentiment Analysis for Comment Response of Video Induced(MSA-CRVI), aims to infer opinions and emotions according to comments response to micro video. Meanwhile, we manually annotate a dataset named Comment Sentiment toward to Micro Video (CSMV) to support this research. It is the largest video multi-modal sentiment dataset in terms of scale and video duration to our knowledge, containing 107, 267 comments and 8, 210 micro videos with a video duration of 68.83 hours. To infer the induced sentiment of comment should leverage the video content, we propose the Video Content-aware Comment Sentiment Analysis (VC-CSA) method as a baseline to address the challenges inherent in this new task. Extensive experiments demonstrate that our method is showing significant improvements over other established baselines. We make the dataset and source code publicly available at https://github.com/IEIT-AGI/MSA-CRVI.
... For precise sentiment inference from comments, it is essential to integrate the video content. Current approaches tend to treat comment sentiment analysis to be a simple NLP task and neglect the semantic connection between videos and comments [42,39,30,10,2,24]. ...
Existing video multi-modal sentiment analysis mainly focuses on the sentiment expression of people within the video, yet often neglects the induced sentiment of viewers while watching the videos. Induced sentiment of viewers is essential for inferring the public response to videos, has broad application in analyzing public societal sentiment, effectiveness of advertising and other areas. The micro videos and the related comments provide a rich application scenario for viewers induced sentiment analysis. In light of this, we introduces a novel research task, Multi-modal Sentiment Analysis for Comment Response of Video Induced(MSA-CRVI), aims to inferring opinions and emotions according to the comments response to micro video. Meanwhile, we manually annotate a dataset named Comment Sentiment toward to Micro Video (CSMV) to support this research. It is the largest video multi-modal sentiment dataset in terms of scale and video duration to our knowledge, containing 107,267 comments and 8,210 micro videos with a video duration of 68.83 hours. To infer the induced sentiment of comment should leverage the video content, so we propose the Video Content-aware Comment Sentiment Analysis (VC-CSA) method as baseline to address the challenges inherent in this new task. Extensive experiments demonstrate that our method is showing significant improvements over other established baselines.
... Several studies, for example [6], [7], and [8], have been performed techniques to analyze opinions posted by users about a particular video. The analysis of user comments as a source can be integrated into several applications such as comment filtering, personal recommendation, and user profiling. ...
Video-sharing sites platforms like YouTube have unique architecture and atmosphere. The comments section is one of the evolution that attracted the users towards expressing opinions and sharing more about videos. Opinions can be used to examine knowledge, user behavior analysis and provide the creator with more ideas to create videos. This paper proposed a novel NLP framework to examine user comments on YouTube and use sentiment analysis to create a short video from positive comments. The results of this study suggest that the framework could be effective to promote the original video using classified community comments. In addition, the results of our implementation indicate that such as framework can be integrated to detect some comments on YouTube and remove negative comments before even posting them.
... SVM, Naive Bayes, Random Forest and Maximum Entropy are used to eliminate the unwanted data and consider those data which yields better results to the user search queries. Smitashree Choudhury et al. in [8] proposed an unsupervised lexicon approach to find the polarity of the user comment. Senti Wordnet is used to find the sentiment polarity and list is prepared by negating the words in the comment. ...
p align="justify">Revolution in social media has attracted the users towards video sharing sites like YouTube. It is the most popular social media site where people view, share and interact by commenting on the videos. There are various types of videos that are shared by the users like songs, movie trailers, news, entertainment etc. Nowadays the most trending videos is the unboxing videos and in particular unboxing of mobile phones which gets more views, likes/dislikes and comments. Analyzing the comments of the mobile unboxing videos provides the opinion of the viewers towards the mobile phone. Studying the sentiment expressed in these comments show if the mobile phone is getting positive or negative feedback. A Hybrid approach combining the lexicon approach Sentiment VADER and machine learning algorithm Naive Bayes is applied on the comments to predict the sentiment. Sentiment VADER has a good impact on the Naive Bayes classifier in predicting the sentiment of the comment. The classifier achieves an accuracy of 79.78% and F1 score of 83.72%.</p
... There have been a number of studies addressing the sentiment of YouTube comments [14][15][16][17]. These methods generally have had different aims and followed several methodologies. ...
Ever since its development in 2005, YouTube has been providing a vital social media platform for video sharing. Unfortunately, YouTube users may have malicious intentions, such as disseminating malware and profanity. One way to do so is using the comment field for this purpose. Although YouTube provides a built-in tool for spam control, yet it is insufficient for combating malicious and spam contents within the comments. In this paper, a comparative study of the common filtering techniques used for YouTube comment spam is conducted. The study deploys datasets extracted from YouTube using its Data API. According to the obtained results, high filtering accuracy (more than 98%) can be achieved with low-complexity algorithms, implying the possibility of developing a suitable browser extension to alleviate comment spam on YouTube in future.
... Many English language sentiment analysis researchers benefited from the wide spread of YouTube and performed their analyses using YouTube publicly available data. Some researchers studied only YouTube comments' sentiments for specific video categories [22]. The authors in [23] collected around 6 million comments for all YouTube video categories, and examined the dependency between the comments' ratings and their sentiments. ...
Abstract—With the current level of ubiquity of social media websites, obtaining the users preferences automatically became a crucial task to assess their tendencies and behaviors online.
Arabic language as one of the most spoken languages in the world and the fastest growing language on the Internet motivates us to provide reliable automated tools that can perform sentiment analysis to reveal users opinions.
In this paper, we present our work of Arabic comments classification based on our collected and manually annotated YouTube Arabic comments. We share our classification results utilizing the most commonly used supervised classifiers: SVMRBF, KNN, and Bernoulli NB classifiers. Experiments were performed using both raw and language-normalized datasets. We show that SVM-RBF outperformed other
classification methods with an f-measure of 88.8% using a normalized dataset with two polarities.
... Hence, there is a need for an automatic content analysis that that can listen to, read and extract relevant information that it is looking for which is termed as 'metadata'. There are many web platforms that are used to share non-textual content such as videos, images and animations that allow users to add comments for each item [1]. Sentiment analysis or opinion mining is one of the great accomplishments of the last decade in the field of Language Technologies. ...
Considerable efforts are currently underway to mitigate the negative impacts of echo chambers, such as increased susceptibility to fake news and resistance towards accepting scientific evidence. Prior research has presented the development of computer systems that support the consumption of news information from diverse political perspectives to mitigate the echo chamber effect. However, existing studies still lack the ability to effectively support the key processes of news information consumption and quantitatively identify a political stance towards the information. In this paper, we present HearHere, an AI-based web system designed to help users accommodate information and opinions from diverse perspectives. HearHere facilitates the key processes of news information consumption through two visualizations. Visualization 1 provides political news with quantitative political stance information, derived from our graph-based political classification model, and users can experience diverse perspectives (Hear). Visualization 2 allows users to express their opinions on specific political issues in a comment form and observe the position of their own opinions relative to pro-liberal and pro-conservative comments presented on a map interface (Here). Through a user study with 94 participants, we demonstrate the feasibility of HearHere in supporting the consumption of information from various perspectives. Our findings highlight the importance of providing political stance information and quantifying users' political status as a means to mitigate political polarization. In addition, we propose design implications for system development, including the consideration of demographics such as political interest and providing users with initiatives.
Ever since its development in 2005, YouTube has been providing a vital social media platform for video sharing. Unfortunately, YouTube users may have malicious intentions, such as disseminating malware and profanity. One way to do so is using the comment field for this purpose. Although YouTube provides a built-in tool for spam control, yet it is insufficient for combating malicious and spam contents within the comments. In this work, we\ have evaluated several top-performance classification techniques\ for such purpose. The statistical analysis of results indicates that, with 99.9% of Naive Bayes, KNN, SVMs are statistically equivalent. Based on this, we have also offered the Tube Spam an accurate online system to filter comments posted on YouTube.
Ever since its development in 2005, YouTube has been providing a vital social media platform for video sharing. Unfortunately, YouTube users may have malicious intentions, such as disseminating malware and profanity. One way to do so is using the comment field for this purpose. Although YouTube provides a built-in tool for spam control, yet it is insufficient for combating malicious and spam contents within the comments. In this work, we\have evaluated several top-performance classification techniques\for such purpose. The statistical analysis of results indicates that, with 99.9% of Naive Bayes, KNN, SVMs are statistically equivalent. Based on this, we have also offered the Tube Spam an accurate online system to filter comments posted on YouTube.
Nowadays, extracting information from social media is providing knowledge about the market and current trends. This paper highlights a novel idea of allowing users to access all such information. A Website is created for visualization of obtained sentiments. It allows users to integrate different social media sites on a single platform for sentimental analysis by providing multiple dashboards under the user’s profile. Users have a choice to provide input from YouTube and/or Twitter to get sentiments. Users need to provide any URL/hashtag. Once input is provided, a trained LSTM model classifies the sentiments as positive, negative, and neutral. Dashboard creation process gets initiated and is deployed on cloud. Classified sentiments are analyzed and are reflected on the dashboard. The dashboard consists of positive, negative, and neutral sentiment’s count, line graph, and pie chart on a real-time basis. These results provide a better understanding of the market and people’s opinion. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
It cannot be denied that YouTube is now the most popular video sharing website. Opinions written from viewers could be an important asset for the development of a company or YouTubers. In this report, the technique used to analyze those comments is called sentiment analysis or opinion mining techniques. Although this knowledge or technique is not new, it is still important to continue studying this, given the importance of the viewer comments. This paper is the result of a literature review of several studies that have been conducted using various methods in sentiment analysis. The purpose of this analytical study is to obtain a theoretical basis that can support further research by studying the working methods, advantages, and disadvantages of each method. For the reason of having no knowledge on how the method works will affect the results and will be a waste of time. The results of comparisons from research that have been done, showing that the Naïve bayes algorithm has a higher accuracy, then SVM then DT. But this is a preliminary result as no study has used all methods at once in a case.
Sentiment analysis or opinion mining is the field of study related to analyze opinions, sentiments, evaluations, attitudes, and emotions of users which they express on social media and other online resources. The revolution of social media sites has also attracted the users towards video sharing sites, such as YouTube. The online users express their opinions or sentiments on the videos that they watch on such sites. This project presents a brief survey of techniques to analyze opinions posted by users about a particular video. Opinion mining or comments toward attitude evaluation, individual entity, are usually called sentiment. Everyone is free to give opinion related with the present opinions on youtube. Hence people have a free will to express their opinion regarding the performance. Due to the raise of many critics that appear in a short amount of time, there a need to conduct a analysis on opinion mining. The process of searching or tracing the natural language to find patterns or moods of society against certain products, people or topics is called Sentiment Analysis. Sentiment analysis is also often referred to as the opinion of mining.[1] 1 The sentiment analysis has received considerable attention since the research of Pang, Turney, Goldberg and Zhu. Sentiment analysis techniques can support many decisions in many scenarios. This study uses three class attributes, which are positive, neutral and negative, because in the internet the comments that appear can be positive, neutral and negative comments.[2] 2 TextBlob is a Python (2 and 3) library for processing textual data. It provides a simple API for diving into common natural language processing (NLP) tasks such as part-of-speech tagging, noun phrase extraction, sentiment analysis, classification, translation, and more. To get this this library, follow the command.
The recent boom in social networks usage has generated some multilingual opinion data for low-resource languages. Luganda is one of the major languages in Uganda, thus it is a low-resource language and Luganda corpora for sentiment analysis especially for YouTube is not easily available. In this paper, we propose assumptions to guide collection of Luganda comments using Luganda YouTube video opinions for sentiment analysis. We evaluate the suitability of our clean YouTube comments (158) dataset for sentiment analysis using selected machine learning and deep learning classification algorithms. Given the low-resource setting, the dataset performs best with Gaussian Naive Bayes for machine learning (55%) and deep learning Multilayer Perceptron sequential model scoring (68.8%) when dataset splitting is at 10% for test set with Luganda comment segmentation.
ResearchGate has not been able to resolve any references for this publication.