Emerging Paradigms in Machine Learning

Volume 13 of the series Smart Innovation, Systems and Technologies pp 215-247

Rough Non-deterministic Information Analysis: Foundations and Its Perspective in Machine Learning

  • Hiroshi SakaiAffiliated withDepartment of Basic Sciences, Faculty of Engineering, Kyushu Institute of Technology Email author 
  • , Hitomi OkumaAffiliated withFaculty of Education and Welfare Science, Oita University
  • , Michinori NakataAffiliated withFaculty of Management and Information Science, Josai International University

* Final gross prices may vary according to local VAT.

Get Access


This chapter focuses on a mathematical framework for handling information incompleteness, which is deeply related to machine learning. Recently, the handling of the information incompleteness in data sets is recognized to be very important research area for machine learning. We have already proposed a framework Rough Non − deterministic Information Analysis (RNIA). This is a rough sets based framework for handling not only definite (or complete) information but also indefinite (or incomplete) information. This RNIA handles lots of aspects in tables with the information incompleteness, i.e., rough sets based issues, data dependencies, question-answering, rule generation, estimation of actual values, etc. Each aspect is extended from tables with complete information to tables with incomplete information according to the modal concepts. We survey this RNIA, and we describe the perspective of RNIA with respect to machine learning.