Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2021
|
| Online Access: | https://doi.org/10.5281/zenodo.15692470 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p><a href="https://ijetrm.com/issues/files/Jun-2021-18-1750268819-JUNE202127.pdf" target="_blank" rel="noopener">Real-time analytics </a>is increasingly critical for timely, informed decision-making across industries. This paper<br>examines how integrating cloud-based event streaming platforms with machine learning enables real-time<br>predictive analytics at scale. We design an architecture leveraging Amazon Kinesis and Apache Kafka for highthroughput data ingestion, combined with AWS Lambda and SageMaker to perform instant predictive inference<br>on streaming data. Key challenges of scalability, fault tolerance, and low latency are addressed through cloudmanaged services and distributed processing techniques. Experimental results demonstrate that our Kinesis-based<br>pipeline handled over one million events per minute with sub-second end-to-end latency, and real-time processing<br>improved predictive accuracy by 20% compared to batch methods. Apache Kafka achieved comparable<br>performance but required more manual tuning to scale. These findings underscore that cloud event streaming,<br>when properly architected, can significantly reduce latency and enhance the accuracy of predictive models during<br>live data processing. The outcomes align with prior research on scalable cloud analytics and illustrate practical<br>strategies for deploying real-time predictive analytics across domains.</p>