International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 05 | May 2022
p-ISSN: 2395-0072
www.irjet.net
Popularity Prediction of video content over cloud-based CDNs using end-user interest Burhanuddin Kanorwala1, Siddharth Kothari2, Saad Karol3, Prof. Rupali Satpute 1,2,3Student,
Dept. Electronics & Telecommunications, K. J. Somaiya Institute Of Engineering & Technology, Sion, Mumbai, India Assistant Professor, Dept. Electronics & Telecommunications, K. J. Somaiya Institute Of Engineering &Technology, Sion, Mumbai, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - More is typically thought to be better,
CDN cache servers by using machine learning approaches to predict
however, this is not the case when it comes to data. Every second, nearly a million minutes of video content crosses the network, and it will take an individual over 5 million years to watch all the videos across the Internet each month. We can't keep up with the tremendous growth of internet data because our capacity to handle, filter, and manage it isn't keeping up. CDNs have grown increasingly popular in today's world as the demand for faster and more consistent data access grows. However, storing all of the content on CDN servers has grown challenging. As a result, the goal of this project is to optimize video material stored in CDNs. In this project, we present a push-based caching technique for improving end-user quality of experience by locating relevant popular films in accordance with an area. Toclassifyfilmsaslow, medium, or very popular, a semisupervised machine learning strategy was used. Popularity prediction of web content has exploded in popularity in recent years. However, there has been minimal research on forecasting video popularity based on pre-existing and significant video attributes. The experimental findings indicate high accuracy, validating the parameter choices and related processing. Key Words: Popularity prediction, Video content, CDNs
video popularity. Prediction is the process of predicting the likelihood of a certain result on a specific dataset once an algorithm has been trained on it. For each tuple in the new data, the machine learning algorithm creates probable values for some unknown variable, allowing the model to determine what the most likely value of the unknown variable is for that case. Although current algorithms have been demonstrated to be useful for predicting video popularity, there are still a few unanswered concerns. One of the most important considerations is how much data is needed to train the model. More user behavior data or video correlation elements in general, according to research, contribute to greater algorithm performance.[9] Several YouTube measurement studies have looked at various statistical and behavioral features, but none have looked at popularity as thoroughly as we did. Recent works, [1]–[6] study distribution and temporal patterns of view count. Because consumers' attention and time are limited, it's not unusual that online video popularity is frequently distributed asymmetrically: a small number of videos acquire the bulk of views, while the rest of the videos are scarcely observed. [7] [8].
1. INTRODUCTION
2. Methodology
A Content Delivery Network (CDN) is a collection of server nodes located around the world that are used to download resources (typically static content such as photos and JavaScript), resulting in faster delivery and lower latency. CDNs have grown in popularity throughout the years. Recent improvements in utility and cloud computing have made it possible to lease resources such as storage and bandwidth to create cloud-based Content Delivery Networks (CDNs). The petabytes of user-generated content traffic streaming over CDNs demonstrate the vastness of the content collection, making it critical to detect and select popular content for storage on the surrogate servers. The ability to predict video popularity is critical for the design and assessment of a CDN. The purpose of this research is to improve the effectiveness of
To research, review and find and implement various methodologies for predicting the popularity of content to be put in cloud CDNs to
© 2022, IRJET
|
Impact Factor value: 7.529
|
Determine how the popularity of video content evolves based on this evolution description
Seeking to estimate the popularity of video content in the future, by considering the current information available.
Reduce the Caching server load.
Increase hit ratio of content solicitation.
Improve user experience by reducing response time. We have developed a three-phase approach.
ISO 9001:2008 Certified Journal
|
Page 1234