Adaptive Learning for Service Monitoring Data

TL;DR

This paper presents an adaptive classification method using Learn++ for real-time service monitoring data, capable of handling evolving data distributions and concept drift to improve accuracy.

Contribution

It introduces an incremental learning approach that updates models continuously and manages concept drift in industrial service monitoring applications.

Findings

Effective handling of concept drift in real-time data

Improved classification accuracy over static models

Demonstrated on industrial monitoring data

Abstract

Service monitoring applications continuously produce data to monitor their availability. Hence, it is critical to classify incoming data in real-time and accurately. For this purpose, our study develops an adaptive classification approach using Learn++ that can handle evolving data distributions. This approach sequentially predicts and updates the monitoring model with new data, gradually forgets past knowledge and identifies sudden concept drift. We employ consecutive data chunks obtained from an industrial application to evaluate the performance of the predictors incrementally.