Updating knowledge in multigranulation decision‐theoretic rough set model based on decision support degree. Issue 13 (28th July 2020)

• Record Type: Journal Article
• Authors: Lin, Guoping; Liu, Fengling; Chen, Shengyu; Yu, Xiaolong
• Is Part Of: Journal of engineering. Volume 2020:Issue 13(2020)
• Page Start: 335
• Page End: 343
• Publication Date: 2020-07-28
• Subjects: Bayes methods -- granular computing -- decision theory -- rough set theory -- learning (artificial intelligence); multigranulation decision‐theoretic rough set model -- decision support degree -- decision risk minimisation -- Bayesian decision‐theoretic -- incremental learning -- static algorithm; Engineering -- Periodicals; Engineering; Electronic journals; Periodicals; 620.005
• DOI: 10.1049/joe.2019.1192
• Languages: English
• ISSNs: 2051-3305

Abstract

Abstract : Based on the majority rules, a multigranulation decision‐theoretic rough set model based on the decision support degree is proposed, in which the thresholds can be computed by the decision risk minimisation based on the Bayesian decision‐theoretic. In various practical situations, information systems may alter dynamically with time. Incremental learning is an alternative manner for maintaining knowledge by utilising previous computational results under dynamic data. Therefore, the authors investigate dynamic approaches to update the knowledge in the new model when adding or deleting granular structures. Besides, the corresponding dynamic and static algorithms are designed and their time complexities are analysed. Finally, comparative experiments by using six data sets from UCI are carried out; the results illustrate that the proposed dynamic algorithm is effective and is more efficient than the static algorithm.