A practical outlier detection approach for mixed-attribute data
作者：
Highlights：
• A principled approach to identify outliers in mixed-attribute data is proposed.
• It is based on novel outlier scoring functions and utilizes the beta mixture model.
• It is able to automatically discriminate outliers from inliers.
• It can be applied to both mixed-type attribute and single-type attribute data.
• It achieves competitive results in comparison to mainstream approaches.
摘要
•A principled approach to identify outliers in mixed-attribute data is proposed.•It is based on novel outlier scoring functions and utilizes the beta mixture model.•It is able to automatically discriminate outliers from inliers.•It can be applied to both mixed-type attribute and single-type attribute data.•It achieves competitive results in comparison to mainstream approaches.
论文关键词：Data mining,Outlier detection,Mixed-attribute data,Mixture model,Bivariate beta
论文评审过程：Received 23 March 2015, Revised 11 July 2015, Accepted 12 July 2015, Available online 22 July 2015, Version of Record 29 August 2015.
论文官网地址：https://doi.org/10.1016/j.eswa.2015.07.018
