Assessing partially ordered clustering in a multicriteria comparative context
作者：
Highlights：
• Data, which are characterized by indicators or criteria with a preference for either small or large values, are not specifically processed in pattern recognition.
• Concepts from the Multicriteria Decision Aid Analysis (MCDA) field are useful to take preference characteristics into account.
• Preference, dominance and Pareto frontiers are useful concepts to analyze data partitions in a multicriteria context.
• For clustering problems, crossover between pattern recognition and MCDA approaches make it possible to identify relationships between clusters that are partially ordered.
• Pattern recognition and MCDA offer complementary approaches to data analysis that should be shared by researchers from these two communities.
摘要
•Data, which are characterized by indicators or criteria with a preference for either small or large values, are not specifically processed in pattern recognition.•Concepts from the Multicriteria Decision Aid Analysis (MCDA) field are useful to take preference characteristics into account.•Preference, dominance and Pareto frontiers are useful concepts to analyze data partitions in a multicriteria context.•For clustering problems, crossover between pattern recognition and MCDA approaches make it possible to identify relationships between clusters that are partially ordered.•Pattern recognition and MCDA offer complementary approaches to data analysis that should be shared by researchers from these two communities.
论文关键词：Clustering,K-means,Multicriteria,Partial ordering,Partition,Preference,Quality assessment
论文评审过程：Received 13 May 2020, Revised 6 November 2020, Accepted 22 January 2021, Available online 29 January 2021, Version of Record 13 February 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2021.107850
