Adaptive learning with covariate shift-detection for non-stationary environments

Raza, H; Prasad, G; Li, Y

Abstract

Learning with dataset shift is a major challenge in non-stationary environments wherein the input data distribution may shift over time. Detecting the dataset shift point in the time-series data, where the distribution of time-series shifts its properties, is of utmost interest. Dataset shift exists in a broad range of real-world systems. In such systems, there is a need for continuous monitoring of the process behavior and tracking the state of the shift so as to decide about initiating adaptation in a timely manner. This paper presents an adaptive learning algorithm with dataset shift-detection using an exponential weighted moving average (EWMA) model based test in a non-stationary environment. The proposed method initiates the adaptation by reconfiguring the knowledge-base of the classifier. This algorithm is suitable for real-time learning in non-stationary environments. Its performance is evaluated through experiments using synthetic datasets. Results show that it reacts well to different covariate shifts.

Citation

Raza, H., Prasad, G., & Li, Y. (2014, September). Adaptive learning with covariate shift-detection for non-stationary environments. Presented at 14th UK Workshop on Computational Intelligence (UKCI2014), Bradford, England

• Presentation Conference Type: Other
• Conference Name: 14th UK Workshop on Computational Intelligence (UKCI2014)
• Conference Location: Bradford, England
• Start Date: Sep 8, 2014
• End Date: Sep 10, 2014
• Online Publication Date: Oct 20, 2014
• Publication Date: Sep 10, 2014
• Deposit Date: Jun 19, 2015
• Publisher: Institute of Electrical and Electronics Engineers
• Book Title: 2014 14th UK Workshop on Computational Intelligence (UKCI)
• DOI: https://doi.org/10.1109/UKCI.2014.6930161
• Publisher URL: http://dx.doi.org/10.1109/UKCI.2014.6930161
• Related Public URLs: http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=6917611; http://www.computing.brad.ac.uk/ukci2014/
• Additional Information: Event Type : Workshop