An online deep extreme learning machine based on forgetting mechanism

• Main Author: Liu Buzhong
• Format: Article
• Language: zho
• Published: National Computer System Engineering Research Institute of China 2018-07-01
• Series: Dianzi Jishu Yingyong
• Subjects: neural networks; feature learning; extreme learning machine; forgotten mechanism; deep autoencoder
• Online Access: http://www.chinaaet.com/article/3000087137

The development of deep learning promotes the development of deep online learning, and online learning tends to have strong effectiveness. Based on the principle of online extreme learning machine and the principle of autoencoder of deep extreme learning machine, an unsupervised online deep extreme learning machine autoencoder is proposed. The forgotten mechanism is introduced into the online deep ELM-AE, proposing the FOS-DELM-AE. With FOS-DELM-AE do unsupervised feature learning,and FOS-ELM do supervised target learning, OS-ELM has deep network structure，and with the ability of online sequential learning. Finally, the effectiveness of FOS-ELM algorithm is verified by experiments based on RMSE and R-square.