Extracting boolean and probabilistic rules from trained neural networks

• 劉, 鵬宇

  Bioinformatics Center, Institute for Chemical Research, Kyoto University

• 阿久津, 達也

  Department of Computer Science, Ben-Gurion University of the Negev

• Akutsu, Tatsuya

  Bioinformatics Center, Institute for Chemical Research, Kyoto University

抄録

This paper presents two approaches to extracting rules from a trained neural network consisting of linear threshold functions. The first one leads to an algorithm that extracts rules in the form of Boolean functions. Compared with an existing one, this algorithm outputs much more concise rules if the threshold functions correspond to 1-decision lists, majority functions, or certain combinations of these. The second one extracts probabilistic rules representing relations between some of the input variables and the output using a dynamic programming algorithm. The algorithm runs in pseudo-polynomial time if each hidden layer has a constant number of neurons. We demonstrate the effectiveness of these two approaches by computational experiments.

収録刊行物

• Neural Networks

  Neural Networks 126 300-311, 2020-06

  Elsevier BV

キーワード

• Neural networks
• Boolean functions
• Rule extraction
• Dynamic programming

詳細情報

• CRID: 1050569159935526784
• NII論文ID: 120006936364
• ISSN: 08936080
• DOI: 10.1016/j.neunet.2020.03.024
• HANDLE: 2433/259350
• Web Site: https://api.elsevier.com/content/article/PII:S0893608020301180?httpAccept=text/xml
• Web Site: https://api.elsevier.com/content/article/PII:S0893608020301180?httpAccept=text/plain
• 本文言語コード: en
• 資料種別: journal article
• データソース種別:
  • IRDB
  • Crossref
  • CiNii Articles
  • KAKEN