ENHANCED CONCEPT DRIFT IN PROCESS MINING

Computer Engineering

Economic success of an organization is dependent on how they react & adapt changes in its operating environment. Research aims to enhance the existing Drift Detection with Change process discovery in complex datasets. Configurable process model describes a family of similar process models. Process variants discovered using concept drift can be merged to derive a configurable process model. Online learning algorithms often have to operate in the presence of concept drifts. Recent study revealed that different diversity levels in an ensemble of learning machines are required to maintain high generalization on both old & new concepts. Based on this study of diversity with different strategies to deal with drifts, we propose new online ensemble learning approach called Diversity for Dealing with Drifts (DDD). DDD maintains ensembles with different diversity levels and is able to achieve better accuracy than other approaches. It is very robust & outperforming in terms of accuracy when there are false positive drift detections.

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