Research Commons

Batch-incremental versus instance-incremental learning in dynamic and evolving data

Research Commons

An outage of Research Commons required for upgrades will occur from 5:00-5:30pm Monday 25/05/2015. Research Commons (including the thesis deposit form) will be unavailable to all users during this outage.

Show simple item record Read, Jesse Bifet, Albert Pfahringer, Bernhard Holmes, Geoffrey 2012-11-18T22:37:41Z 2012-11-18T22:37:41Z 2012 2012
dc.identifier.citation Read, J., Bifet, A., Pfahringer, B. & Holmes, G. (2012). Batch-incremental versus instance-incremental learning in dynamic and evolving data. In 11th International Symposium on Intelligent Data Analysis, IDA 2012, Helsinki, October 25-27 2012, (pp. 313-323). Helsinki, Finland. en_NZ
dc.description.abstract Many real world problems involve the challenging context of data streams, where classifiers must be incremental: able to learn from a theoretically- infinite stream of examples using limited time and memory, while being able to predict at any point. Two approaches dominate the literature: batch-incremental methods that gather examples in batches to train models; and instance-incremental methods that learn from each example as it arrives. Typically, papers in the literature choose one of these approaches, but provide insufficient evidence or references to justify their choice. We provide a first in-depth analysis comparing both approaches, including how they adapt to concept drift, and an extensive empirical study to compare several different versions of each approach. Our results reveal the respective advantages and disadvantages of the methods, which we discuss in detail. en_NZ
dc.language.iso en
dc.publisher Springer en_NZ
dc.subject data streams en_NZ
dc.subject dynamic en_NZ
dc.subject evolving en_NZ
dc.subject incremental en_NZ
dc.subject on-line en_NZ
dc.title Batch-incremental versus instance-incremental learning in dynamic and evolving data en_NZ
dc.type Conference Contribution en_NZ
dc.identifier.doi 10.1007/978-3-642-34156-4_29 en_NZ
pubs.elements-id 22694

Full-text options:

This item appears in the following Collection(s)

Show simple item record

Search Research Commons

Advanced Search



About Research Commons

My Account

Usage Statistics


  • Bookmark and Share