Binary classification in unstructured space with hypergraph case-based reasoning
作者：
Highlights：
• A new method for binary classification in unstructured space is proposed.
• The method is well calibrated: its support can directly be interpreted as probability.
• Without data assumption or ad-hoc metric, it competes with more specialized methods.
• It ranks second after Neural Network compared to 9 other methods on 7 datasets.
• It gives consistently good results without hyperparameter tuning or feature engineering.
摘要
•A new method for binary classification in unstructured space is proposed.•The method is well calibrated: its support can directly be interpreted as probability.•Without data assumption or ad-hoc metric, it competes with more specialized methods.•It ranks second after Neural Network compared to 9 other methods on 7 datasets.•It gives consistently good results without hyperparameter tuning or feature engineering.
论文关键词：Binary classification,Hypergraph,Case-based reasoning
论文评审过程：Received 11 June 2018, Revised 26 February 2019, Accepted 7 March 2019, Available online 18 March 2019, Version of Record 13 July 2019.
论文官网地址：https://doi.org/10.1016/j.is.2019.03.005
