GAN-Based Prediction of Time Series

被引：0

作者：

Festag, Sven ^{[1
]}

Spreckelsen, Cord ^{[1
]}

机构：

[1] Jena Univ Hosp, Inst Med Stat, Comp & Data Sci, Jena, Germany

来源：

PUBLIC HEALTH AND INFORMATICS, PROCEEDINGS OF MIE 2021 | 2021年 / 281卷

关键词：

GAN; time series; patient care; imputation; forecasting;

D O I：

10.3233/SHTI210214

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

The study aims at generating initial and directional insights in the applicability of conditional recurrent generative adversarial nets for the imputation and forecasting of medical time series data. Our experiment with blood pressure series showed that a generative recurrent autoencoder exhibits significant individual learning progress but needs further tuning to benefit from joint training.

引用

页码：500 / 501

页数：2