2001年
On-line learning methods for Gaussian processes
ARTIFICIAL NEURAL NETWORKS-ICANN 2001, PROCEEDINGS
- ,
- ,
- 巻
- 2130
- 号
- 開始ページ
- 292
- 終了ページ
- 299
- 記述言語
- 英語
- 掲載種別
- DOI
- 10.1007/3-540-44668-0_42
- 出版者・発行元
- SPRINGER-VERLAG BERLIN
This article proposes two modifications of Gaussian processes, which aim to deal with dynamic environments. One is a weight decay method that gradually forgets the old data, and the other is a time stamp method that regards the time course of data as a Gaussian process. We show experimental results when these modifications are applied to regression problems in dynamic environments.
- リンク情報
- ID情報
-
- DOI : 10.1007/3-540-44668-0_42
- ISSN : 0302-9743
- Web of Science ID : WOS:000173024600041