YouTube Comments Research Articles

YOUTUBE COMMENT ANALYSIS USING MACHINE LEARNING

Analyzing comments on YouTube may yield invaluable insights on the demographics, hobbies, and preferences of the viewership. By looking into comment patterns, we may identify recurring themes, intriguing topics, and even potential collaborations. We may group comments using natural language processing techniques into three categories: good, negative, and neutral. This research helps identify key trends, gauge viewer sentiment, and assess how well a video is received overall. Material providers may use the information gathered from comments to better tailor their material to the specific needs and tastes of their audience, which will ultimately boost viewer satisfaction and engagement. KEY WORDS: Machine learning, Bidirectional Encoder Representation from Transformers (BERT), YouTube comments, Transformers.

Exploring YouTube Comments to Understand Public Sentiment on COVID-19 Vaccines through Deep Learning-based Sentiment Analysis

COVID-19 was first found in China in 2019. Since then, it has quickly spread around the world, which has led to a lot of news stories and social media posts about the pandemic. YouTube, a popular video-sharing website, has become a valuable source of information on COVID-19 and other topics. However, it can be difficult to extract useful insights from the vast array of user comments that accompany these videos. One potential method for understanding public sentiment is to use sentiment analysis, which involves classifying text as positive, negative, or neutral. In this study, the dataset of over 44,000 YouTube comments related to COVID-19 vaccines was used, which was filtered to a total of 16,073 comments for analysis. The data was cleaned and organised using NeatText and then processed using GloVe word embedding, a technique for establishing statistical relationships between words. Based on the experiment, the performances of three different types of deep learning techniques: recurrent neural networks (RNN), gated recurrent units (GRU) and long short-term memory (LSTM) are compared in accurately classifying the sentiment of the comments. The study found that the GRU had the highest accuracy of 80.19%, followed by the LSTM with 79.00% accuracy, and the RNN with 67.15% accuracy.

Ondel-Ondel Culture: People's Perception on the Shifting in Function and Value Degradation

Ondel-ondel is one of Jakarta’s most prominent cultures. It is a giant doll made of woods and decorated with cotton clothes and human-like embellishments. However, there has been a shift in the way people perceive its role. Ondel-Ondel, which was initially used in significant ceremonial contexts, has evolved into a tool for busking or street entertainment, or more commonly known as Ondel-Ondel Ngamen. At the moment, Ondel-Ondel Ngamen has captured big interest of online users on YouTube. There is a variety of comments on YouTube regarding the existence and representation of Ondel-ondel Ngamen. In this study, the author aims to evaluate YouTube comments regarding people's perceptions of the Ondel-Ondel culture. The author applies a qualitative approach by providing the data from the top-three Indonesian news-related YouTube channels; CNN Indonesia, KompasTV, and Metrotvnews. The dataset is taken from any other scholarly articles that support this analysis. In this study, the author utilizes Stuart Hall's encoding-decoding and commodification theories to evaluate how people perceive Ondel-Ondel, which has undergone a shift in function and value degradation. This study's findings indicate that Ondel-Ondel is vulnerable to commodification. It is no longer regarded as a sacred aspect of rituals or noteworthy occasions.

Youtube Comment Feature Selection And Classification Using Fused Machine Learning

The exponential rise of internet platforms, notably YouTube, has resulted in a massive volume of user-generated material, including video comments. Understanding audience input and enhancing user experience require analyzing and forecasting the mood of YouTube comments. This work provides a comprehensive method for YouTube comment prediction and sentiment classification that combines feature selection using Recursive Feature Elimination (RFE), Elastic Net Random Forest with Logistic Regression (RF with LR), and Principal Component Analysis (PCA). The first stage is to choose the most informative features from a given dataset using RFE, a common approach for doing so. RFE aids in the elimination of unnecessary or redundant features, resulting in enhanced model performance and decreased computing complexity. The Elastic Net Random Forest with Logistic Regression (RF with LR) technique is then used to construct a robust sentiment classification model. Elastic Net regularization combines the advantages of both L1 (Lasso) and L2 (Ridge) regularization, allowing for improved feature selection and management of multicollinearity concerns. By integrating many decision trees, the Random Forest ensemble approach improves the model's predictive potential even more. We employ Principal Component Analysis (PCA) to increase the classification model's effectiveness and solve possible difficulties created by high-dimensional data. PCA decreases the dataset's complexity while retaining its fundamental qualities, resulting in a more manageable and efficient feature space for classification. Finally, we compare the performance of three prominent classifiers on the preprocessed dataset: Linear Support Vector Machine (LSVM), Gaussian Naive Bayes (GNB), Logistic Regression (LR), and Decision Tree (DT). We can select the best-performing model for YouTube comment categorization by comparing these classifiers.

YouTube discourse of the Oting massacre in Nagaland: investigating affiliations, sentiments and Naga identity negotiation in YouTube comments

Personae identification and understanding the social bonds shared in online interactions are found to be key to combating misinformation and conspiratorial discourse as well as aiding in developing solutions for specific communities. The 2021 Oting massacre in Nagaland that killed 14 civilians sparked public outrage and widespread protests calling for justice and the repeal of the Armed Forces (Special Powers) Act in northeast India. The massacre received a lot of attention, including YouTube videos, and reignited discussion on militarization, Naga identity and a backlash over the botched operation. A close qualitative analysis of 500 randomly extracted comments from 10 selected YouTube videos identified 15 personae with unique linguistic patterns that reveal specific ideational targets and their affiliation strategies in the discourse. A sentiment analysis of 11,294 comments shows a higher negative score linked to condemnation, hate speech and conspiracy than a positive score linked to solidarity and empathy. ‘Nagaism’ as being a Naga is reimagined and reinforced via YouTube, countering misconceptions about Naga identity.

Informal Learning Trends on YouTube during the COVID‐19 Pandemic: A Topic Modeling Analysis

ABSTRACTThis study explores the impact of the COVID‐19 pandemic on informal online learning behavior through the lens of online video tutorials hosted on YouTube. Over 40,000 YouTube comments were collected for topic modeling analysis and uncovered thirteen latent topics related to three learning behavior types, that is, environment structuring, help‐seeking, and self‐evaluation. These learning behaviors underwent varying degrees of change after the pandemic declaration and normalized after that, highlighting the importance of adapting teaching methods to meet the changing needs of learners in response to any learning disruptions.

Web-ресурстардағы экстремисттік мәліметтерді анықтауға арналған машиналық әдістерді оқыту және сынау үшін қазақ тіліндегі мәтіндер корпусын құру

The article deals with the issue of creating a corpus in the Kazakh language, which is necessary to create a semantic model for identifying extremist information on web resources. As a result of the study, text data was collected from closed groups of the VKontakte social network, YouTube comments and news sites using previously developed analytics. The texts were classified into two classes: «extremist» and «neutral». A corpus for further training of models was created. The total number of extremist messages included in the corpus is about 1,200, and the total number of words in the corpus is about 140,000 words. The distribution of extremist and neutral texts was carried out, the analysis of the corpus was carried out using a word cloud. The Python 3.7 programming language with pandas, numpy, matplotlib, plotly, bokeh, cufflinks, spacy, googletrans packages was used as the main calculation and visualization libraries for data analysis. Signs of texts of extremist content were identified.

Appraising Unmet Needs and Misinformation Spread About Polycystic Ovary Syndrome in 85,872 YouTube Comments Over 12 Years: Big Data Infodemiology Study.

Polycystic ovary syndrome (PCOS) is the most common endocrinopathy in women, resulting in substantial burden related to metabolic, reproductive, and psychological complications. While attempts have been made to understand the themes and sentiments of the public regarding PCOS at the local and regional levels, no study has explored worldwide views, mainly due to financial and logistical limitations. YouTube is one of the largest sources of health-related information, where many visitors share their views as questions or comments. These can be used as a surrogate to understand the public's perceptions. We analyzed the comments of all videos related to PCOS published on YouTube from May 2011 to April 2023 and identified trends over time in the comments, their context, associated themes, gender-based differences, and underlying sentiments. After extracting all the comments using the YouTube application programming interface, we contextually studied the keywords and analyzed gender differences using the Benjamini-Hochberg procedure. We applied a multidimensional approach to analyzing the content via association mining using Mozdeh. We performed network analysis to study associated themes using the Fruchterman-Reingold algorithm and then manually screened the comments for content analysis. The sentiments associated with YouTube comments were analyzed using SentiStrength. A total of 85,872 comments from 940 PCOS videos on YouTube were extracted. We identified a specific gender for 13,106 comments. Of these, 1506 were matched to male users (11.5%), and 11,601 comments to female users (88.5%). Keywords including diagnosing PCOS, symptoms of PCOS, pills for PCOS (medication), and pregnancy were significantly associated with female users. Keywords such as herbal treatment, natural treatment, curing PCOS, and online searches were significantly associated with male users. The key themes associated with female users were symptoms of PCOS, positive personal experiences (themes such as helpful and love), negative personal experiences (fatigue and pain), motherhood (infertility and trying to conceive), self-diagnosis, and use of professional terminology detailing their journey. The key themes associated with male users were misinformation regarding the "cure" for PCOS, using natural and herbal remedies to cure PCOS, fake testimonies from spammers selling their courses and consultations, finding treatment for PCOS, and sharing perspectives of female family members. The overall average positive sentiment was 1.6651 (95% CI 1.6593-1.6709), and the average negative sentiment was 1.4742 (95% CI 1.4683-1.4802) with a net positive difference of 0.1909. There may be a disparity in views on PCOS between women and men, with the latter associated with non-evidence-based approaches and misinformation. The improving sentiment noticed with YouTube comments may reflect better health care services. Prioritizing and promoting evidence-based care and disseminating pragmatic online coverage is warranted to improve public sentiment and limit misinformation spread.

»The perfect film for posers.«

The article deals with the aesthetics and critical fan reception of the Black Metal biopic lords of chaos (GB/SE 2018, Jonas Åkerlund) along the dimension of its ›authenticity‹. In the film itself and in paratextual discourses in the form of exemplary YouTube comments, divergent concepts of authenticity are demonstrated and the difficulty (up to the point of impossibility) discussed to satisfy a normative and identity-creating idea of ›trueness‹ for Black Metal fans cinematically. The aim is to demonstrate the importance of reception for the film, which challenges the notion of an ideal-typical reader.

Analysis of YouTube Comments to Inform the Design of Virtual Reality Training Simulations to Target Emotional Arousal

Analysis of YouTube Comments to Inform the Design of Virtual Reality Training Simulations to Target Emotional Arousal

From YouTube Comments to Insights: A Sentiment Analysis of Opinions on Productivity Tools

Abstract: YouTube, a diverse content hub, acts as a prominent platform for both viewing videos and engaging through comments that express opinions. By carefully analyzing these comments, sentiments associated with products or topics become evident. This research delves into sentiment analysis of YouTube comments, specifically focusing on sentiments towards productivity tools and technologies. Through this examination of conveyed sentiments, it deepens understanding of public perceptions and attitudes toward these tools. Utilizing YouTube's reach and engagement, the study systematically collects comment data using the YouTube API and employs the established VADER sentiment analysis framework. Employing a Bag-of-Words approach, the research incorporates machine learning algorithms like Naive Bayes and Random Forest, achieving accuracy rates of 84.01% and 91.34%, respectively. In-depth temporal analysis of user engagement patterns uncovers trends, with heightened engagement followed by decline, correlating with external events, notably the Covid-19 pandemic. These insights enhance sentiment analysis and illuminate dynamics between societal occurrences, user sentiments, and digital dialogues, offering perspectives on evolving opinions about productivity tools and technologies

The Politics of Quarantine: The Political Implications of Negative Social Affects in the Early Stage of COVID-19 in South Korea

We analyze the interrelationships between quarantine-related events and negative social affects during COVID-19 in South Korea. We explain the impact of these relationships on the political nature of quarantine. We conducted a discourse analysis of YouTube comments posted on COVID-19-related videos during 2020. We analyzed these affects and their social and cognitive triggers using cognition-emotion theory and affect theory. Contingent COVID-19 outbreaks in South Korea and ambiguous information contributed to fear and anger. Social disgust combined with anger as mass infection events occurred over time, influencing the attitudes of South Korean citizens who support the country’s social distancing policy. We find that social disgust realized through exclusion and purification is associated with the politics of quarantine. Consequently, we show that the political disgust surrounding quarantine was embedded within social disgust. This study paves the path for further research on the governmentality of affects and mobility in the context of epidemics.

Multilingual hope speech detection: A Robust framework using transfer learning of fine-tuning RoBERTa model

Hope Speech Detection (HSD) from social media is a new direction for promoting and supporting positive content to encourage harmony and positivity in society. As users of social media belong to different linguistic communities, hope speech detection is rarely studied as a multilingual task considering low-resource languages. Moreover, prior studies explored only monolingual techniques, and the Russian language is not addressed. This study tackles the issue of Multi-lingual Hope Speech Detection (MHSD) in English and Russian languages using the transfer learning paradigm with fine-tuning approach. We explore joint multi-lingual and translation-based approaches to tackle the task of multilingualism, where the latter approach adopts the translation mechanism to transform all content into one language and then classify them. The joint multi-lingual method handles it by designing a universal classifier for various languages. We explore the strengths of the Robustly Optimized BERT Pre-Training Approach (RoBERTa) that showed a benchmark in capturing the semantics and contextual information within the content. The proposed framework consists of several stages: 1) data preprocessing, 2) representation of data using RoBERTa models, 3) fine-tuning phase, and 4) classification of hope speech into two labels. A new Russian corpus for hope speech detection is built, containing YouTube comments. Several experiments are conducted in English and Russian languages by using semi-supervised bilingual English and Russian datasets. The findings show that the proposed framework demonstrated benchmark performance and outperformed the baselines. Furthermore, the translation-based approach (Russian-RoBERTa) offered the best performance by achieving 94% accuracy and 80.24% f1-score.

YouTube comments on violin instruction videos: an analysis of comments in Turkish

ABSTRACT This research examines the comments on violin instruction videos on YouTube in the context of practice challenges encountered by learners while playing the violin and solutions provided by instructors to these challenges. 205 comments related to the research context were included in the analysis among a total of 4934 comments from 398 violin instruction videos in Turkish and they were examined through qualitative content analysis. It has been observed that all of the practice challenges experienced by learners while playing the violin are at the beginner level. Furthermore, instructors offer solutions to these challenges by providing short instructions, mainly through the comments section. They also emphasise patience and perseverance for learners. In some cases, it has been noticed that instructors assist learners by suggesting off-site communication, encouraging them to watch instructors’ videos again, and recommending taking private lessons to address their specific problems.

YouTube Comment Scraping and Sentiment Analysis Using Python and Machine Learning

YouTube has become a massive platform for video sharing and discussion, with millions of users expressing their thoughts and opinions through comments. For content producers, marketers, and academics, gleaning significant insights from these comments can yield useful data. This project explores the combination of machine learning and web scraping techniques to scrape YouTube comments and perform sentiment analysis. By leveraging machine learning algorithms, we aim to analyze sentiments, opinions, and trends within the vast amount of textual data generated on the platform. The process involves collecting comments using the YouTube Data API or web scraping libraries, followed by preprocessing steps to clean and transform the data. Next, Machine learning methods like Naive Bayes, SVM, or deep learning models are used to train sentiment analysis models. These approaches include Bag-of-Words, TF-IDF, and word embedding. Our ability to comprehend the overall emotion conveyed by viewers is subsequently made possible by the trained models, which are then utilized to anticipate the sentiment of fresh comments. Additionally, the ability to identify very favorable or negative remarks is made possible by sorting comments based on sentiment scores. It is essential to consider the complexities of language and contextual nuances when conducting sentiment analysis. For many players in the YouTube ecosystem, this technique does provide useful insights into audience feedback, content strategy, and data-driven decision-making.

Analisis Sentimen Pengguna Youtube Mengenai Analog Switch Off Menggunakan Word Embedding Dan Metode Long Short-Term

Analog Switch Off (ASO) or migration program from analog television to digital television is a program issued by the Ministry of Communication and Informatics in Indonesia. Some people provide different responses and opinions on YouTube comments about ASO. There are those who give positive or neutral comments. However, there were also those who gave negative comments. Sentiment analysis is a process that is carried out automatically in studying, retrieving, and processing textual data to obtain information and see responses or opinions about an issue or object towards positive, neutral or negative opinions. Thus sentiment analysis can be used as a reference in making organizational decisions, improving a service, or as a review of a product. Sentiment analysis was performed using word embedding with Word2Vec, and sentiment classification using the Long Short-Term Memory (LSTM) method. The results of the evaluation test are 92% accuracy, 92% precision, 92% recall, and 92% f1-score.

The usage of Sinographs in relational communication about ethnic Chinese identities on YouTube

Abstract This study analyzes the usage of Sinographs (Chinese characters) across globally dispersed user communities on YouTube, and focuses on their usage patterns in relational communication about ethnic Chinese identities. This is achieved by analyzing YouTube comments on the official music video for the Mandopop (Mandarin pop music) song “Fragile” (玻璃心) as a corpus. Firstly, the history and media environment of ethnic Chinese communities are reviewed to provide context for why “Fragile” sparked heated discussion in the YouTube comment section. Then, the methodological approach of corpus pragmatics, the data, and the methodology including corpus analytical process and tools used are discussed. Finally, the analysis results are presented in two parts: the usage of sentence patterns that communicate relationships and identities, and the usage of metamessages that indicate ethnic Chinese identities.

Virality is not enough when the audience misses the point: A Content Analysis of YouTube Comments on Environmental Documentary “Sexy Killers”

During the Indonesian presidential election campaign, a documentary on coal mining problems in Indonesia titled “ Sexy Killers” became viral and was highly engaged on YouTube in early 2019, attracting 36 million views and above 100,000 comments. The documentary also received mainstream media coverage and had such a far-reaching appeal that it swayed many voters into not voting in the election. Showing the impact of coal mining on Indonesia’s environment and its people, the documentary exposed networks of business owners and politicians who had stakes in the coal mining company. This research thus delves into understanding the public discourse, especially the high number of YouTube comments on the documentary. The analysis revealed that despite the documentary’s widespread popularity, the discussion within the comments did not focus on the environmental problems (72.9%). Most of the comments discussed the political campaign and presidential candidates (41.3%), instead of discussing the environmental impact of coal mining in Indonesia (13.9%). The research is valuable in highlighting the centrality of political debates dwarfing real environmental issues in Indonesia. Environment activists use documentary movies as one of their advocacy strategies, but it must be followed by audience engagement strategy to assure the discussions are still relevant to increase audience awareness on the issue. Keywords: Social media, environment, text analysis, YouTube, Indonesia.

Look Who’s Watching: Platform Labels and User Engagement on State-Backed Media Outlets

Recently, social media platforms have introduced several measures to counter misleading information. Among these measures are “state-media labels” which help users identify and evaluate the credibility of state-backed news. YouTube was the first platform to introduce labels that provide information about state-backed news channels. While previous work has examined the efficiency of information labels in controlled lab settings, few studies have examined how state-media labels affect users’ perceptions of content from state-backed outlets. This article proposes new methodological and theoretical approaches to investigate the effect of state-media labels on users’ engagement with content. Drawing on a content analysis of 8,071 YouTube comments posted before and after the labeling of five state-funded channels (Al Jazeera English [AJE], China Global Television Network, Russia Today [RT], TRT World, and Voice of America [VOA] News), this article analyses the effect that YouTube’s labels had on users’ engagement with state-backed media content. We found that the labels had no impact on the amount of likes that the videos received before and after the policy introduction, except for RT which received less likes after it was labeled. However, for RT, comments left by users were associated with 30% decrease in the likelihood of observing a critical comment following the policy implementation, and a 70% decrease in likelihood of observing a critical comment about RT as a media source. While other state-funded broadcasters, like AJE and VOA News, received fewer critical comments after YouTube introduced its policy; this relationship was associated with how political the video was, rather than the policy change. Our study contributes to the ongoing discussion on the efficacy of platform governance in relation to state-backed media, showing that audience preferences impact the effectiveness of labels.

The expression of hate speech against Afro-descendant, Roma, and LGBTQ+ communities in YouTube comments

Abstract This paper addresses the specificities of online hate speech against the Afro-descendant, Roma, and LGBTQ+ communities in Portugal. The research is based on the analysis of CO-HATE, a corpus composed of 20,590 YouTube comments, which were manually annotated following detailed guidelines created for that purpose. We applied methods from corpus linguistics to assess the prevalence of overt and covert hate speech, counter-speech, and offensive speech, considering different grounds of discrimination, and to investigate the main linguistic and rhetorical strategies underlying hatred messages. The research results highlight the importance of tackling covert hate speech, a recurring phenomenon often anchored in irony and fallacious argumentation, including the emotional appeal to fear and the implicit call to action. We believe this study will aid in advancing the analysis of online hate speech, while promoting the development of efficient automated detection models, specifically regarding the Portuguese language.