An embedded imputation method via Attribute-based Decision Graphs
作者：
Highlights：
• Attribute-based Decision Graphs represent the correlation among data attributes.
• Similar data instances induce similar subgraphs in the AbDG.
• Imputation partially matches instances to the AbDG searching for a proper subgraph.
• The method has low computational costs and handles high rates of missing values.
• Results show the method is efficient to impute data prior to classification tasks.
摘要
•Attribute-based Decision Graphs represent the correlation among data attributes.•Similar data instances induce similar subgraphs in the AbDG.•Imputation partially matches instances to the AbDG searching for a proper subgraph.•The method has low computational costs and handles high rates of missing values.•Results show the method is efficient to impute data prior to classification tasks.
论文关键词：Missing attribute value,Data imputation,Single imputation,Attribute-based Decision Graphs,Machine learning based imputation Methods
论文评审过程：Received 25 June 2015, Revised 15 March 2016, Accepted 16 March 2016, Available online 25 March 2016, Version of Record 6 April 2016.
论文官网地址：https://doi.org/10.1016/j.eswa.2016.03.027
