Presentation 2015-03-06
An Efficient Incremental-Decremental Data Filtering Method for Stream Learning
Kota MATSUI, Saori KURUMI, Ichiro TAKEUCHI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we propose an efficient data filtering method for learning problems with streaming data. Our problem setup contains two major features. First one is the finite-buffer setting, that is, data set which constructs a classifier is restricted to have a finite size. In this case, we have to consider both how to add a new data point to the current data set, and how to exclude a data point from that. We especially consider the former, that is, data addition. Second, we also consider the outlier filtering. It is important to discard outlier from the training data set for guaranteeing quality of the classifier. We use the validation error as a filter criterion and show that our proposal can efficiently (from the viewpoint of both) update the training data set.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) stream learning / data filtering / finite-buffer / stream aware criterion / ball test
Paper # IBISML2014-94
Date of Issue

Conference Information
Committee IBISML
Conference Date 2015/2/26(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Information-Based Induction Sciences and Machine Learning (IBISML)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An Efficient Incremental-Decremental Data Filtering Method for Stream Learning
Sub Title (in English)
Keyword(1) stream learning
Keyword(2) data filtering
Keyword(3) finite-buffer
Keyword(4) stream aware criterion
Keyword(5) ball test
1st Author's Name Kota MATSUI
1st Author's Affiliation Nagoya Institute of Technology()
2nd Author's Name Saori KURUMI
2nd Author's Affiliation Nagoya Institute of Technology
3rd Author's Name Ichiro TAKEUCHI
3rd Author's Affiliation Nagoya Institute of Technology
Date 2015-03-06
Paper # IBISML2014-94
Volume (vol) vol.114
Number (no) 502
Page pp.pp.-
#Pages 7
Date of Issue