Cartesian genetic programming for diagnosis of Parkinson disease through handwriting analysis: Performance vs. interpretability issues
作者：
Highlights：
• A comparison among AI classification methods for the diagnosis of PD is performed.
• Classification is performed on handwriting samples belonging to benchmark datasets.
• Cartesian genetic programming (CGP) outperforms white box approaches in accuracy
• CGP outperforms black box methods in interpretability by providing explicit rules.
• CGP classification rules provide guidelines for the design of diagnostic protocols.
摘要
•A comparison among AI classification methods for the diagnosis of PD is performed.•Classification is performed on handwriting samples belonging to benchmark datasets.•Cartesian genetic programming (CGP) outperforms white box approaches in accuracy•CGP outperforms black box methods in interpretability by providing explicit rules.•CGP classification rules provide guidelines for the design of diagnostic protocols.
论文关键词：Explainable artificial intelligence,Parkinson disease,Evolutionary computation
论文评审过程：Received 18 November 2019, Revised 16 September 2020, Accepted 3 November 2020, Available online 10 November 2020, Version of Record 19 November 2020.
论文官网地址：https://doi.org/10.1016/j.artmed.2020.101984
