A Mechanism for Supporting the Peak/Trough Detection in the Concept Drifting Environment

Abstract

The primary objective of this research endeavor is to tackle the difficult challenge of identifying the elusive aspect peak/trough of the business cycle, which is known to exhibit the tendency of concept drift. To effectively confront this issue, we have put forth a supervised learning approach referred to as PLM E. This approach encompasses the integration of the envelope concept and the sequence-based sliding window mechanism along with the pupil learning mechanism (PLM), aimed at effectively detecting the elusive aspect peak/trough of the business cycle within changing data patterns. To confirm the effectiveness of our proposed PLM E approach, this study has conducted some experiments utilizing a real-world copper price forecasting dataset.