Data imputation and compression for Parkinson's disease clinical questionnaires

作者：

Highlights：

• Proposed autoencoder is efficient for incomplete medical questionnaires compression.

• Autoencoders impute data better than linear methods, regardless of amount missing.

• A low number of components is necessary for data imputation through compression.

• Data with a high number of missing values are informative and shouldn’t be discarded.

摘要

•Proposed autoencoder is efficient for incomplete medical questionnaires compression.•Autoencoders impute data better than linear methods, regardless of amount missing.•A low number of components is necessary for data imputation through compression.•Data with a high number of missing values are informative and shouldn’t be discarded.

论文关键词：Autoencoders,Medical questionnaires,Data imputation,Parkinson's disease,PPMI

论文评审过程：Received 13 September 2019, Revised 27 January 2021, Accepted 21 February 2021, Available online 5 March 2021, Version of Record 11 March 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102051