Explaining data with descriptions

作者:

Highlights:

• Data descriptions are compact, readable and insightful formulas of boolean predicates that represent datasets;

• Generating the optimal data description is NP-hard and depends on user needs;

• A dynamic programming approach generates descriptions at interactive speed;

• Experiments against real datasets show that descriptions support users in getting insight from the data.

摘要

•Data descriptions are compact, readable and insightful formulas of boolean predicates that represent datasets;•Generating the optimal data description is NP-hard and depends on user needs;•A dynamic programming approach generates descriptions at interactive speed;•Experiments against real datasets show that descriptions support users in getting insight from the data.

论文关键词:Data explanation,Data exploration,Outlier analysis,Data profiling

论文评审过程:Received 25 September 2019, Revised 20 April 2020, Accepted 26 April 2020, Available online 4 May 2020, Version of Record 21 May 2020.

论文官网地址:https://doi.org/10.1016/j.is.2020.101549