Engagement and Popularity Dynamics of YouTube Videos and Sensitivity to Meta-Data

Hoiles, William; Aprem, Anup; Krishnamurthy, Vikram

doi:10.1109/tkde.2017.2682858

Cited by 52 publications

(54 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Being a social media, Youtube also presents social network data associated with videos; such data have been used to understand and sometimes predict video traffic patterns and to improve the effectiveness of caching strategies. Other works such as [29] and [30] analyzed Youtube data in order to understand users' engagement dynamics. Social dynamics and video metadata (title, tag, thumbnail, and description) have been shown to influence, and thus be correlated with, the number of views [29], the number of likes per view, and sentiment feedback in the comments [30].…”

Section: Content Update Statisticsmentioning

confidence: 99%

Analysis of Data Persistence in Collaborative Content Creation Systems: The Wikipedia Case

et al. 2019

View full text Add to dashboard Cite

A very common problem in designing caching/prefetching systems, distribution networks, search engines, and web-crawlers is determining how long a given content lasts before being updated, i.e., its update frequency. Indeed, while some content is not frequently updated (e.g., videos), in other cases revisions periodically invalidate contents. In this work, we present an analysis of Wikipedia, currently the 5th most visited website in the world, evaluating the statistics of updates of its pages and their relationship with page view statistics. We discovered that the number of updates of a page follows a lognormal distribution. We provide fitting parameters as well as a goodness of fit analysis, showing the statistical significance of the model to describe the empirical data. We perform an analysis of the views-updates relationship, showing that in a time period of a month, there is a lack of evident correlation between the most updated pages and the most viewed pages. However, observing specific pages, we show that there is a strong correlation between the peaks of views and updates, and we find that in more than 50% of cases, the time difference between the two peaks is less than a week. This reflects the underlying process whereby an event causes both an update and a visit peak that occurs with different time delays. This behavior can pave the way for predictive traffic analysis applications based on content update statistics. Finally, we show how the model can be used to evaluate the performance of an in-network caching scenario.

show abstract

Section: Content Update Statisticsmentioning

confidence: 99%

Analysis of Data Persistence in Collaborative Content Creation Systems: The Wikipedia Case

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Insight into the dynamics of social sensors in YouTube can be used to predict how users will interact with posted video content. These results are important for designing methods for optimizing user engagement and for improving the efficiency of content distribution networks [87,164,88]. Estimating the popularity of YouTube videos based on meta-level features is a challenging problem given the diversity of users and content providers.…”

Section: Social Interaction Of Youtube Consumersmentioning

confidence: 99%

“…The classifier was able to achieve a classification accuracy of 66%. In [87] visual perception and extreme learning machines were applied to the meta-level features of videos and found to be able to accurately estimate the (popular/unpopular) videos with an accuracy of 80%. It was determined that the main meta-level features that impact video engagement include: first day view count , number of subscribers, and contrast of the video thumbnail.…”

Section: Social Interaction Of Youtube Consumersmentioning

confidence: 99%

Dynamics of Information Diffusion and Social Sensing

Krishnamurthy

Hoiles²

2018

Cooperative and Graph Signal Processing

Self Cite

View full text Add to dashboard Cite

Statistical inference using social sensors is an area that has witnessed remarkable progress in the last decade. It is relevant in a variety of applications including localizing events for targeted advertising, marketing, localization of natural disasters and predicting sentiment of investors in financial markets. This chapter presents a tutorial description of four important aspects of sensing-based information diffusion in social networks from a communications/signal processing perspective. First, diffusion models for information exchange in large scale social networks together with social sensing via social media networks such as Twitter is considered. Second, Bayesian social learning models and risk averse social learning is considered with applications in finance and online reputation systems. Third, the principle of revealed preferences arising in micro-economics theory is used to parse datasets to determine if social sensors are utility maximizers and then determine their utility functions. Finally, the interaction of social sensors with YouTube channel owners is studied using time series analysis methods. All four topics are explained in the context of actual experimental datasets from health networks, social media and psychological experiments. Also, algorithms are given that exploit the above models to infer underlying events based on social sensing. The overview, insights, models and algorithms presented in this chapter stem from recent developments in network science, economics and signal processing. At a deeper level, this chapter considers mean field dynamics of networks, risk averse Bayesian social learning filtering and quickest change detection, data incest in decision making over a directed acyclic graph of social sensors, inverse optimization problems for utility function estimation (revealed preferences) and statistical modeling of interacting social sensors in YouTube social networks.

show abstract

“…The performance of the risk-neutral and riskaverse caching policies are evaluated using real-world data from the YouTube social network in Sec. 6. The results show that a 6% reduction in the average delay can be achieved if the uncertainty of the content requests is accounted for, and a 60% reduction in average delay is achieved if both the uncertainty and femtocell routing protocol are accounted for compared to the risk-neutral caching policy that neglects the routing protocol.…”

Section: Introductionmentioning

confidence: 98%

“…Content feature based methods use the uploader, textual, and image features of the content to predict the content requests. These methods include multivariate linear regression [4], Markov clustering [5], and extreme learning machines [6]. All these methods provide point forecasts (expected value) for the content requests.…”

Section: Introductionmentioning

confidence: 99%

Risk-Averse Caching Policies for YouTube Content in Femtocell Networks using Density Forecasting

Hoiles

Tanzil

Krishnamurthy

2021

IEEE Trans. Cloud Comput.

Self Cite

View full text Add to dashboard Cite

The paper presents risk-neutral and risk-averse caching policies that can be deployed in a femtocell network with limited storage capacity to reduce the time delay of servicing content requests. The caching policies use a forecasting algorithm to estimate the cumulative distribution function of content requests based on the content features. Given the cumulative distribution function, a mixed-integer linear program is used to compute where to cache content in the femtocell network. The caching policies account for the uncertainty associated with estimating the content requests using the coherent Conditional Value-at-Risk (CVaR) measure. For a large number of content, a risk-neutral caching policy is constructed that accounts for both the content features and routing protocol that only requires the evaluation of a unimodular linear program. Using data from YouTube (comprising 25,000 videos) and the NS-3 simulator, the caching policies reduce the delay of retrieving content in femtocell networks compared with industry standard caching policies. Specifically, a 6% reduction in delay is achieved by accounting for the uncertainty, and a 60% reduction in delay is achieved if both the uncertainty and femtocell routing protocol are accounted for compared to the risk-neutral caching policy that neglects the routing protocol.

show abstract

Engagement and Popularity Dynamics of YouTube Videos and Sensitivity to Meta-Data

Cited by 52 publications

References 25 publications

Analysis of Data Persistence in Collaborative Content Creation Systems: The Wikipedia Case

Analysis of Data Persistence in Collaborative Content Creation Systems: The Wikipedia Case

Dynamics of Information Diffusion and Social Sensing

Risk-Averse Caching Policies for YouTube Content in Femtocell Networks using Density Forecasting

Contact Info

Product

Resources

About