論文

査読有り
2020年

Unsupervised Feature Value Selection Based on Explainability.

Agents and Artificial Intelligence
  • Kilho Shin
  • ,
  • Kenta Okumoto
  • ,
  • David Lawrence Shepard
  • ,
  • Akira Kusaba
  • ,
  • Takako Hashimoto
  • ,
  • Jorge Amari
  • ,
  • Keisuke Murota
  • ,
  • Junnosuke Takai
  • ,
  • Tetsuji Kuboyama
  • ,
  • Hiroaki Ohshima

12613 LNAI
開始ページ
421
終了ページ
444
記述言語
掲載種別
研究論文(国際会議プロシーディングス)
DOI
10.1007/978-3-030-71158-0_20
出版者・発行元
Springer

The problem of feature selection has been an area of considerable research in machine learning. Feature selection is known to be particularly difficult in unsupervised learning because different subgroups of features can yield useful insights into the same dataset. In other words, many theoretically-right answers may exist for the same problem. Furthermore, designing algorithms for unsupervised feature selection is technically harder than designing algorithms for supervised feature selection because unsupervised feature selection algorithms cannot be guided by class labels. As a result, previous work attempts to discover intrinsic structures of data with heavy computation such as matrix decomposition, and require significant time to find even a single solution. This paper proposes a novel algorithm, named Explainability-based Unsupervised Feature Value Selection (EUFVS), which enables a paradigm shift in feature selection, and solves all of these problems. EUFVS requires only a few tens of milliseconds for datasets with thousands of features and instances, allowing the generation of a large number of possible solutions and select the solution with the best fit. Another important advantage of EUFVS is that it selects feature values instead of features, which can better explain phenomena in data than features. EUFVS enables a paradigm shift in feature selection. This paper explains its theoretical advantage, and also shows its applications in real experiments. In our experiments with labeled datasets, EUFVS found feature value sets that explain labels, and also detected useful relationships between feature value sets not detectable from given class labels.

リンク情報
DOI
https://doi.org/10.1007/978-3-030-71158-0_20
DBLP
https://dblp.uni-trier.de/rec/conf/icaart/ShinOSKHAMTKO20
URL
https://dblp.uni-trier.de/rec/conf/icaart/2020s
URL
https://dblp.uni-trier.de/db/conf/icaart/icaart2020s.html#ShinOSKHAMTKO20
Scopus
https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85103477615&origin=inward
Scopus Citedby
https://www.scopus.com/inward/citedby.uri?partnerID=HzOxMe3b&scp=85103477615&origin=inward
ID情報
  • DOI : 10.1007/978-3-030-71158-0_20
  • ISSN : 0302-9743
  • eISSN : 1611-3349
  • ISBN : 9783030711573
  • DBLP ID : conf/icaart/ShinOSKHAMTKO20
  • SCOPUS ID : 85103477615

エクスポート
BibTeX RIS