Handling data irregularities in classification: Foundations, trends, and future challenges
作者：
Highlights：
• Data irregularities can significantly degrade the performance of classifiers.
• We present a comprehensive taxonomy and survey of various data irregularities.
• We discuss prominent methods to handle distribution and feature-based irregularities.
• We highlight the co-occurrences and interrelations among different irregularities.
• We unearth a number of promising future research avenues.
摘要
•Data irregularities can significantly degrade the performance of classifiers.•We present a comprehensive taxonomy and survey of various data irregularities.•We discuss prominent methods to handle distribution and feature-based irregularities.•We highlight the co-occurrences and interrelations among different irregularities.•We unearth a number of promising future research avenues.
论文关键词：Data irregularities,Class imbalance,Small disjuncts,Class-distribution skew,Missing features,Absent features
论文评审过程：Received 4 October 2017, Revised 16 January 2018, Accepted 4 March 2018, Available online 14 March 2018, Version of Record 24 May 2018.
论文官网地址：https://doi.org/10.1016/j.patcog.2018.03.008
