Articles
| Open Access |
DOI: UTILIZING A SCALABLE AI/ML-BASED DATA ANOMALY DETECTION TOOL TO IMPROVE DATA QUALITY IN VIDEO STREAMING SERVICES
Alexander Motylev , Director, Data Test Engineering at Paramount Global / PlutoTV Miami, Florida, United StatesAbstract
In today's world of online cinemas and streaming platforms such as Netflix and YouTube, data quality plays a key role in ensuring high user satisfaction. A scalable AI/ML anomaly detection tool is used to improve data quality and enhance the reliability of video streams. This paper examines various approaches to anomaly detection, including supervised and unsupervised learning, as well as deep learning methods such as autoencoders and recurrent neural networks. In addition, the application of AI/ML for predicting user behavior and optimizing the resources of video streaming services is analyzed. The introduction of such technologies allows not only to improve the quality of video streams but also to reduce the cost of content moderation and network resource management. The prospects for the development of these technologies and their impact on the video-streaming industry are discussed.
Keywords
video streaming, data anomaly detection tools, AI-based data anomalies
References
How AI simplifies and automates the work of streaming services. [Electronic resource] Access mode: https://www.comnews.ru/content/230555/2023-12-07/2023-w49/1013/kak-ii-uproschaet-i-avtomatiziruet-rabotu-strimingovykh-servisov (accessed 07/12/2024).
Anitha Kumari K, Dharani D. A Survey on Anomaly Detection using Unsupervised Learning Techniques. [Electronic resource] Access mode: https://www.researchgate.net/publication/343318335_A_Survey_on_Anomaly_Detection_using_Unsupervised_Learning_Techniques (accessed 12.07.2024).
Popova I.A. Detection of anomalies in a data set using machine learning algorithms without a teacher isolation forest and local outlier factor // StudNet. 2020. No.12. pp.1460-1470.
Chernikov A.D. Prediction and recognition of objects in a video stream using deep learning // Bulletin of Science. 2023. No.9 (66). pp.209-215.
Video Analysis to Detect Suspicious Activity Based on Deep Learning. [Electronic resource] Access mode: https://dzone.com/articles/video-analysis-to-detect-suspicious-activity-based (accessed 12.07.2024).
Kalytyuk I.S., Frantsuzova G.A., Gunko A.V. On the issue of choosing methods for predictive analysis of social media data // Automation and software Engineering. 2019. No.4 (30). pp.9-17.
User behavior: how to understand and predict user behavior using click modeling. [Electronic resource] Access mode: https://fastercapital.com/ru/content/Поведение-пользователей--как-понять-и-спрогнозировать-поведение-пользователей-с-помощью-моделирования-кликов.html (accessed 07/12/2024).
Denis Khryashchev,Alexandru Papiu, Jiamin Xuan, Olivia Dinica, Kyle Hubert, Huy Vo Who Watches What: Forecasting Viewership for the Top 100 TV Networks // Computational Data and Social Networks. 2019. pp.163-174.
Lyndon Nixon Predicting Your Future Audience's Popular Topics to Optimize TV Content Marketing Success // Conference: Proceedings of the 2nd International Workshop on AI for Smart TV Content Production, Access and Delivery. 2020. p.5.
Sizikov D.O. Time series analysis method and its mathematical model in software // International Scientific Research Journal. 2024. No.3 (141) . [Electronic resource] Access mode: https://research-journal.org/archive/3-141-2024-march/10.23670/IRJ.2024.141.5 (accessed 07/12/2024).
Python | ARIMA Model for Time Series Forecasting. [Electronic resource] Access mode: https://translated.turbopages.org/proxy_u/en-ru.ru.8f7a0cad-66981a60-3df9103b-74722d776562/https/www.geeksforgeeks.org/python-arima-model-for-time-series-forecasting / (accessed 12.07.2024).
Dozhdikov Anton Valentinovich Forecasting the results of film distribution using machine learning // VTE. 2023. No.4. pp. 93-114.
Hybrid Machine Learning Approach to Popularity Prediction of Newly Released Contents for Online Video Streaming Service. [Electronic resource] Access mode: https://translated.turbopages.org/proxy_u/en-ru.ru.859ed0ea-66981acf-db75a817-74722d776562/https/paperswithcode.com/paper/hybrid-machine-learning-approach-to (accessed 12.07.2024).
Morozova V. I. Forecasting by machine learning method // Young Scientist. 2022. No. 21 (416). pp. 202-204.
Kusakina D. S. Development and research of a method for detecting anomalies in a video stream based on the analysis of the context of a scene // Young Scientist. 2023. No. 19 (466). pp. 15-18.
Building architectures for real-time data processing using Apache Kafka, Flink and Druid. [Electronic resource] Access mode: https://habr.com/ru/companies/timeweb/articles/783758 / (accessed 12.07.2024).
Anomaly detection with TensorFlow. [Electronic resource] Access mode: https://translated.turbopages.org/proxy_u/en-ru.ru.41c286df-66981b41-8a53e5ad-74722d776562/https/www.geeksforgeeks.org/anomaly-detection-with-tensorflow / (accessed 12.07.2024).
Orlova I.V. Using the Prophet package in forecasting time series // Fundamental Research. 2021. No. 3. pp. 94-102.
ELK Stack Tutorial: Get Started with Elasticsearch, Logstash, Kibana, & Beats. [Electronic resource] Access mode: https://translated.turbopages.org/proxy_u/en-ru.ru.2192893c-66981b77-65adc37a-74722d776562/https/www.tutorialspoint.com/elk-stack-tutorial-get-started-with-elasticsearch-logstash-kibana-amp-beats (accessed 12.07.2024).
How to use the Kubernetes Executor in Airflow in production. [Electronic resource] Access mode: https://www.restack.io/docs/airflow-kubernetes-executor (accessed 12.07.2024).
Monitoring of WebRTC streams using Prometheus and Grafana. [Electronic resource] Access mode: https://habr.com/ru/companies/flashphoner/articles/568784 / (accessed 07/12/2024).
Prometheus vs Grafana: Top Differences. [Electronic resource] Access mode: https://translated.turbopages.org/proxy_u/en-ru.ru.973f9ca9-66981ba6-c413aef2-74722d776562/https/www.geeksforgeeks.org/prometheus-vs-grafana / (accessed 12.07.2024).
Integration of third-party video analytics. [Electronic resource] Access mode: https://flussonic.ru/doc/integrate-external-video-analytics / (accessed 12.07.2024).
Machine Learning for Anomaly Detection. [Electronic resource] Access mode: https://translated.turbopages.org/proxy_u/en-ru.ru.7c9cebae-66981bc9-0aab5e33-74722d776562/https/www.geeksforgeeks.org/machine-learning-for-anomaly-detection / (accessed 12.07.2024).
Kubernetes for DevOps. [Electronic resource] Access mode: https://github.com/rustamgarifulin/books/blob/master/files/Kubernetes_%D0%B4%D0%BB%D1%8F_DevOps_%D1%80%D0%B0%D0%B7%D0%B2%D0%B5%D1%80%D1%82%D1%8B%D0%B2%D0%B0%D0%BD%D0%B8%D0%B5%2C_%D0%B7%D0%B0%D0%BF%D1%83%D1%81%D0%BA_%D0%B8_%D0%BC%D0%B0%D1%81%D1%88%D1%82%D0%B0%D0%B1%D0%B8%D1%80%D0%BE%D0%B2%D0%B0%D0%BD%D0%B8%D0%B5_%D0%B2.pdf (accessed 12.07.2024).
Immersion in containerization. [Electronic resource] Access mode: https://habr.com/ru/companies/otus/articles/767884 / (accessed 12.07.2024).
CI/CD for Machine Learning: What it is & Benefits in 2024. [Electronic resource] Access mode: https://research.aimultiple.com/ci-cd-machine-learning / (accessed 12.07.2024).
Article Statistics
Copyright License
Copyright (c) 2024 Alexander Motylev
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors retain the copyright of their manuscripts, and all Open Access articles are disseminated under the terms of the Creative Commons Attribution License 4.0 (CC-BY), which licenses unrestricted use, distribution, and reproduction in any medium, provided that the original work is appropriately cited. The use of general descriptive names, trade names, trademarks, and so forth in this publication, even if not specifically identified, does not imply that these names are not protected by the relevant laws and regulations.