CLOUD JOB FAILURE PREDICTION USING AN ADVANCED MULTILAYER VOTING-BASED ENSEMBLE MODEL
DOI:
https://doi.org/10.62643/ijerst.2026.v22.n2.pp215-220Keywords:
Cloud Computing, Job Failure Prediction, Ensemble Learning, Multilayer Voting, Machine Learning, Resource Optimization, Fault ToleranceAbstract
Cloud computing environments have become the backbone of modern distributed applications, supporting large-scale data processing and computational workloads across diverse domains. However, job failures remain a significant challenge, leading to resource wastage, increased operational costs, and degraded system performance. Accurate prediction of job failures can enable proactive resource allocation, fault tolerance, and system optimization. This paper proposes a novel multilayer voting-based framework designed to enhance cloud job failure prediction by integrating multiple machine learning classifiers across hierarchical layers. The framework combines the strengths of diverse models, including decision trees, random forests, gradient boosting, and deep neural networks, through a structured voting mechanism that improves predictive accuracy and robustness. The multilayer design ensures that each layer captures distinct feature interactions, enabling more refined decision-making compared to single-layer or standalone models. Feature engineering techniques, including workload profiling, resource utilization patterns, and historical execution logs, are incorporated to enrich the input dataset. Experimental evaluation is conducted on benchmark cloud workload traces, demonstrating significant improvements in prediction accuracy, precision, recall, and F1-score over traditional methods. The proposed approach reduces false positives and false negatives, ensuring reliable failure prediction in dynamic cloud environments. Furthermore, the framework exhibits scalability and adaptability to varying workload conditions. This study highlights the importance of ensemble learning and layered decision-making in addressing complex prediction problems in cloud computing. The results indicate that the multilayer voting-based framework can serve as an effective tool for enhancing cloud reliability, reducing downtime, and improving overall system efficiency.
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.













