In data stream analysis, detecting the concept drift accurately is important to maintain the classification performance. Most drift detection methods assume that the class labels become available immediately after a data sample arrives. However, it is unrealistic to attempt to acquire all of the labels when processing the data streams, as labeling costs are high and much time is needed. In this paper, we propose a concept drift detection method under the assumption that there is limited access or no access to class labels. The proposed method detects concept drift on unlabeled data streams based on the class label information which is predicted by a classifier or a virtual classifier. Experimental results on synthetic and real streaming data show that the proposed method is competent to detect the concept drift on unlabeled data stream.
Youngin KIM
Agency for Defense Development
Cheong Hee PARK
Chungnam National University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Youngin KIM, Cheong Hee PARK, "An Efficient Concept Drift Detection Method for Streaming Data under Limited Labeling" in IEICE TRANSACTIONS on Information,
vol. E100-D, no. 10, pp. 2537-2546, October 2017, doi: 10.1587/transinf.2017EDP7091.
Abstract: In data stream analysis, detecting the concept drift accurately is important to maintain the classification performance. Most drift detection methods assume that the class labels become available immediately after a data sample arrives. However, it is unrealistic to attempt to acquire all of the labels when processing the data streams, as labeling costs are high and much time is needed. In this paper, we propose a concept drift detection method under the assumption that there is limited access or no access to class labels. The proposed method detects concept drift on unlabeled data streams based on the class label information which is predicted by a classifier or a virtual classifier. Experimental results on synthetic and real streaming data show that the proposed method is competent to detect the concept drift on unlabeled data stream.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2017EDP7091/_p
Copy
@ARTICLE{e100-d_10_2537,
author={Youngin KIM, Cheong Hee PARK, },
journal={IEICE TRANSACTIONS on Information},
title={An Efficient Concept Drift Detection Method for Streaming Data under Limited Labeling},
year={2017},
volume={E100-D},
number={10},
pages={2537-2546},
abstract={In data stream analysis, detecting the concept drift accurately is important to maintain the classification performance. Most drift detection methods assume that the class labels become available immediately after a data sample arrives. However, it is unrealistic to attempt to acquire all of the labels when processing the data streams, as labeling costs are high and much time is needed. In this paper, we propose a concept drift detection method under the assumption that there is limited access or no access to class labels. The proposed method detects concept drift on unlabeled data streams based on the class label information which is predicted by a classifier or a virtual classifier. Experimental results on synthetic and real streaming data show that the proposed method is competent to detect the concept drift on unlabeled data stream.},
keywords={},
doi={10.1587/transinf.2017EDP7091},
ISSN={1745-1361},
month={October},}
Copy
TY - JOUR
TI - An Efficient Concept Drift Detection Method for Streaming Data under Limited Labeling
T2 - IEICE TRANSACTIONS on Information
SP - 2537
EP - 2546
AU - Youngin KIM
AU - Cheong Hee PARK
PY - 2017
DO - 10.1587/transinf.2017EDP7091
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2017
AB - In data stream analysis, detecting the concept drift accurately is important to maintain the classification performance. Most drift detection methods assume that the class labels become available immediately after a data sample arrives. However, it is unrealistic to attempt to acquire all of the labels when processing the data streams, as labeling costs are high and much time is needed. In this paper, we propose a concept drift detection method under the assumption that there is limited access or no access to class labels. The proposed method detects concept drift on unlabeled data streams based on the class label information which is predicted by a classifier or a virtual classifier. Experimental results on synthetic and real streaming data show that the proposed method is competent to detect the concept drift on unlabeled data stream.
ER -