What question did this study set out to answer?

This study evaluates how machine learning can predict system health degradation in SAP ERP systems to mitigate operational risks.

synapse

⌘+K

synapse

⌘+K

June 3, 2026Open Access

Anticipating Failure Before It Happens: A Longitudinal Study of Machine Learning for SAP ERP Performance Forecasting

Key Points

This study evaluates how machine learning can predict system health degradation in SAP ERP systems to mitigate operational risks.
Longitudinal study across seven SAP ERP installations over 24 months in Canada, Brazil, and South Korea.
Evaluation of three different forecasting model architectures including ensemble models.
Utilization of an operational telemetry framework for accurate forecasting.
Ensemble models reduced critical system failures by an average of 64%.
Forecasting accuracy for resource utilization improved by 57%.

Abstract

This paper reports findings from a 24-month longitudinal study across seven SAP ERP installations in Canada, Brazil, and South Korea, evaluating machine learning models for forecasting system health degradation before it causes operational impact. Building on the operational telemetry framework of Veershetty (2026) and deployed within the clean-core extension architecture of Veershetty (2025), the study evaluates three forecasting model architectures and finds that ensemble models reduce critical system failures by an average of 64% and improve resource utilization forecasting accuracy by 57%.

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper